Student Eligibility for a Free Lunch as an SES ... - AERA Homepage
Student Eligibility for a Free Lunch as an SES ... - AERA Homepage
Student Eligibility for a Free Lunch as an SES ... - AERA Homepage
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
<strong>Student</strong> <strong>Eligibility</strong> <strong>for</strong> a <strong>Free</strong> <strong>Lunch</strong> <strong>as</strong> <strong>an</strong><br />
<strong>SES</strong> Me<strong>as</strong>ure in Education Research<br />
Michael Harwell <strong>an</strong>d Br<strong>an</strong>don LeBeau<br />
The use of eligibility <strong>for</strong> a free lunch <strong>as</strong> a me<strong>as</strong>ure of a student’s<br />
socioeconomic status continues to be a fixture of qu<strong>an</strong>titative education<br />
research. Despite its popularity, it is unclear that education<br />
researchers are familiar with what student eligibility <strong>for</strong> a free lunch<br />
does (<strong>an</strong>d does not) represent. The authors examine the National<br />
School <strong>Lunch</strong> Program, which is responsible <strong>for</strong> certifying students <strong>as</strong><br />
eligible <strong>for</strong> a free lunch, <strong>an</strong>d conclude that free lunch eligibility is a<br />
poor me<strong>as</strong>ure of socioeconomic status, which suffers from import<strong>an</strong>t<br />
deficiencies that c<strong>an</strong> bi<strong>as</strong> inferences. A table characterizing key<br />
strengths <strong>an</strong>d weaknesses of variables used <strong>as</strong> me<strong>as</strong>ures of socioeconomic<br />
status is provided to facilitate comparisons.<br />
Keywords: data <strong>an</strong>alysis; me<strong>as</strong>urements; poverty<br />
Analyses of educational data often include student background<br />
variables <strong>as</strong> statistical controls to enh<strong>an</strong>ce the<br />
credibility of inferences. One of the most frequently<br />
used student variables is socioeconomic status (<strong>SES</strong>). The prominence<br />
of <strong>SES</strong> in education is due largely to its widely documented<br />
relationship with achievement, covering more th<strong>an</strong> nine<br />
decades of research (e.g., Bry<strong>an</strong>t, Glazer, H<strong>an</strong>sen, & Kursch,<br />
1974; Colem<strong>an</strong> et al., 1966; Holley, 1916; Lynd & Lynd, 1929;<br />
Sirin, 2005; White, 1982). <strong>SES</strong> is frequently used <strong>as</strong> a covariate<br />
in <strong>an</strong>alyses of educational data (e.g., Dauber, Alex<strong>an</strong>der, &<br />
Entwisle, 1996; Lubienski & Lubienski, 2006; Mathematica<br />
Policy Research, 2008) or <strong>as</strong> a matching variable (e.g., General<br />
Accounting Office, 2003; Pentz et al., 1990) to statistically control<br />
<strong>for</strong> its effects, to incre<strong>as</strong>e statistical power, <strong>an</strong>d to enh<strong>an</strong>ce<br />
causality arguments (White, 1982).<br />
There are m<strong>an</strong>y definitions of <strong>SES</strong>. Three that are representative<br />
are “the social <strong>an</strong>d economic life ch<strong>an</strong>ces individuals experience”<br />
(Powers, 1982, p. 1), “differential access (realized <strong>an</strong>d<br />
potential) to desired resources” (Oakes & Rossi, 2003, p. 775),<br />
<strong>an</strong>d “a shorth<strong>an</strong>d expression <strong>for</strong> variables that enable the placement<br />
of persons, families, households <strong>an</strong>d aggregates such <strong>as</strong> statistical<br />
local are<strong>as</strong>, communities <strong>an</strong>d cities in some hierarchical<br />
order, reflecting their ability to produce <strong>an</strong>d consume the scarce<br />
<strong>an</strong>d valued resources of society” (Hauser & Warren, 1997, p.<br />
178). Life ch<strong>an</strong>ces represent <strong>an</strong> individual’s opportunity to obtain<br />
120<br />
educational ReseaRcheR<br />
the scarce <strong>an</strong>d valued resources of society, which include economic,<br />
social, <strong>an</strong>d cultural resources.<br />
Following the above definitions, a student whose <strong>SES</strong> is “low”<br />
lives in a household in which the householder (e.g., biological<br />
parent, foster parent, guardi<strong>an</strong>) h<strong>as</strong> less income, education, occupational<br />
status (e.g., m<strong>an</strong>ual laborer vs. m<strong>an</strong>ager of a large work<strong>for</strong>ce),<br />
<strong>an</strong>d so <strong>for</strong>th th<strong>an</strong> in the c<strong>as</strong>e of a student whose <strong>SES</strong> is<br />
“high.” As Walpole (2003) points out, “low” <strong>SES</strong> students also<br />
tend to have less access to cultural capital (specialized or insider<br />
knowledge not taught in schools) <strong>an</strong>d social capital (contacts in<br />
networks that c<strong>an</strong> lead to personal or professional gains), which<br />
have been argued to be key components of a student’s educational<br />
success.<br />
An examination of the education research literature shows<br />
that a variety of variables have served <strong>as</strong> <strong>SES</strong> me<strong>as</strong>ures, including<br />
dwelling value neighborhood quality (Montague, 1964); ethnicity<br />
(Burkhead, Fox, & Holl<strong>an</strong>d, 1967); parent income (Worley<br />
& Story, 1967); teacher salaries (Raymond, 1968); parent occupation<br />
(Miner, 1968); student mobility, which is usually related<br />
to tr<strong>an</strong>sferring in or out of a school district during the school year<br />
(Herr & Tobi<strong>as</strong>, 1970); home atmosphere, which typically<br />
reflects householder support of education (Purves, 1973); teacher<br />
estimates of student’s <strong>SES</strong> (Me<strong>an</strong>s, Me<strong>an</strong>s, Osborne, & Elsom,<br />
1973); parent education (St<strong>an</strong>fiel, 1973); number of siblings<br />
(Kerckhoff, 1975); <strong>an</strong>d student eligibility <strong>for</strong> a free or reduced<br />
price lunch (FRL; Stein et al., 2008). <strong>Student</strong>s certified <strong>as</strong> eligible<br />
<strong>for</strong> <strong>an</strong> FRL are distinguished from those not eligible, with the<br />
implication that eligible students have lower <strong>SES</strong>. <strong>Student</strong>s who<br />
receive a free lunch <strong>an</strong>d those who receive a reduced price lunch<br />
will generally not be distinguished. We focus on the use of FRL<br />
<strong>as</strong> a me<strong>as</strong>ure of a student’s <strong>SES</strong> in education research.<br />
Rationale <strong>for</strong> the Current Study<br />
An examination of the FRL variable <strong>as</strong> a me<strong>as</strong>ure of a student’s<br />
<strong>SES</strong> in education research in the United States is timely <strong>for</strong> three<br />
related re<strong>as</strong>ons. First, it is unclear that education researchers are<br />
familiar with what eligibility <strong>for</strong> <strong>an</strong> FRL does (<strong>an</strong>d does not)<br />
represent. Second, the FRL variable possesses several import<strong>an</strong>t<br />
deficiencies that c<strong>an</strong> bi<strong>as</strong> inferences. Third, the continued use of<br />
FRL <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong> in educational studies despite criticism<br />
of this practice (Hauser, 1994; Kurki, Boyle, & Aladjem, 2005)<br />
suggests a need to critically reexamine its use.<br />
We argue that education researchers who use the FRL variable<br />
<strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong> typically do so because this in<strong>for</strong>mation is<br />
Educational Researcher, Vol. 39, No. 2, pp. 120–131<br />
DOI: 10.3102/0013189X10362578<br />
© 2010 <strong>AERA</strong>. http://er.aera.net
e<strong>as</strong>ily accessible <strong>an</strong>d relatively inexpensive. However, the use of<br />
FRL is difficult to defend conceptually, empirically, or methodologically.<br />
Implicit in this argument is that a researcher who carefully<br />
defines what <strong>SES</strong> represents in his or her study is unlikely to<br />
select the FRL variable <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong>.<br />
First we distinguish between poverty <strong>an</strong>d the conceptualization<br />
<strong>an</strong>d me<strong>as</strong>urement of <strong>SES</strong> in education research. Next we<br />
examine the FRL variable <strong>an</strong>d identify characteristics that c<strong>an</strong><br />
bi<strong>as</strong> inferences when this variable serves <strong>as</strong> a me<strong>as</strong>ure of student<br />
<strong>SES</strong>. To provide a deeper underst<strong>an</strong>ding of this variable, we<br />
include a brief history of the National School <strong>Lunch</strong> Program<br />
(NSLP; 2008), which is responsible <strong>for</strong> determining student eligibility<br />
<strong>for</strong> <strong>an</strong> FRL, <strong>an</strong>d its current operation in U.S. schools.<br />
Finally, we provide a table that facilitates comparisons among<br />
variables used <strong>as</strong> <strong>SES</strong> me<strong>as</strong>ures by characterizing key strengths<br />
<strong>an</strong>d weaknesses.<br />
In education research using the FRL variable, it appears that<br />
typically no distinction is drawn between students certified <strong>as</strong><br />
eligible <strong>for</strong> a reduced price lunch (1.59 million) <strong>an</strong>d those certified<br />
<strong>as</strong> eligible <strong>for</strong> a free lunch (16.1 million; NSLP, 2008). For<br />
example, <strong>an</strong>alyses of data collected <strong>for</strong> federally funded education<br />
studies <strong>an</strong>d reports generally make no distinction between these<br />
groups (e.g., Trends in International Mathematics <strong>an</strong>d Science<br />
Study, see Martin, Mullis, & Chrostowski, 2004; National<br />
Center <strong>for</strong> Education Statistics [NCES], 2007).<br />
Poverty <strong>an</strong>d <strong>SES</strong><br />
<strong>SES</strong> <strong>as</strong> typically described in education research differs from poverty,<br />
which is widely used in research are<strong>as</strong> such <strong>as</strong> public health<br />
<strong>an</strong>d the effectiveness of social welfare programs such <strong>as</strong> the Special<br />
Supplemental Nutrition Program <strong>for</strong> Women, Inf<strong>an</strong>ts, <strong>an</strong>d<br />
Children (WIC) <strong>an</strong>d Temporary Assist<strong>an</strong>ce <strong>for</strong> Needy Families<br />
(TANF).<br />
In general, poverty is more narrowly defined th<strong>an</strong> <strong>SES</strong> because it<br />
is almost always linked to the federal government’s poverty levels,<br />
which are strictly income b<strong>as</strong>ed. Official U.S. government poverty<br />
levels first appeared in the mid-1960s (Orsh<strong>an</strong>sky, 1965) <strong>an</strong>d at the<br />
time reflected the price of food <strong>for</strong> a family of a given size times three.<br />
The federal government distinguishes between poverty thresholds,<br />
which are used <strong>for</strong> various statistical purposes such <strong>as</strong> estimating<br />
the number of children in poverty each year, <strong>an</strong>d poverty<br />
guidelines, which represent a simplification of the poverty thresholds<br />
<strong>an</strong>d are issued by the Department of Health <strong>an</strong>d Hum<strong>an</strong><br />
Services to determine eligibility <strong>for</strong> m<strong>an</strong>y federal <strong>as</strong>sist<strong>an</strong>ce programs.<br />
Both me<strong>as</strong>ures rely on household income, which includes<br />
sources such <strong>as</strong> earnings, unemployment compensation, social<br />
security, pension or retirement income, public <strong>as</strong>sist<strong>an</strong>ce, <strong>an</strong>d<br />
dividends. Households with incomes lower th<strong>an</strong> the poverty<br />
guidelines are generally eligible <strong>for</strong> various federal programs such<br />
<strong>as</strong> WIC <strong>an</strong>d <strong>an</strong> FRL through the NSLP.<br />
<strong>SES</strong>, on the other h<strong>an</strong>d, is widely used in education research<br />
<strong>an</strong>d is often <strong>as</strong>sessed with me<strong>as</strong>ures that vary in their connection<br />
to income, such <strong>as</strong> householder occupation <strong>an</strong>d FRL eligibility.<br />
For example, Blau <strong>an</strong>d Dunc<strong>an</strong> (1967) described a conceptual<br />
model of the impact of householder <strong>SES</strong> on achievement, Spaeth<br />
(1976) offered a theory of how <strong>SES</strong> affects a student’s cognitive<br />
development in the home, <strong>an</strong>d Adler et al. (1994) described the<br />
effects of <strong>SES</strong> on health.<br />
The import<strong>an</strong>ce of <strong>SES</strong> in research <strong>an</strong>d policy recently led the<br />
Americ<strong>an</strong> Psychological Association (APA) to commission a t<strong>as</strong>k<br />
<strong>for</strong>ce on <strong>SES</strong>. This t<strong>as</strong>k <strong>for</strong>ce produced a report that operationally<br />
defined the scope, nature, r<strong>an</strong>ge, parameters, <strong>an</strong>d effects of socioeconomic<br />
inequalities in the United States <strong>an</strong>d recommended<br />
mech<strong>an</strong>isms <strong>an</strong>d structures that would more effectively address,<br />
on <strong>an</strong> <strong>as</strong>sociation-wide b<strong>as</strong>is, the causes <strong>an</strong>d the impact of socioeconomic<br />
inequality (APA, 2007b). APA also created the Office<br />
on Socioeconomic Status, which is “responsible <strong>for</strong> directing,<br />
overseeing, facilitating <strong>an</strong>d promoting psychology’s contribution<br />
to the underst<strong>an</strong>ding of <strong>SES</strong> <strong>an</strong>d the lives <strong>an</strong>d well-being of the<br />
poor” (APA, 2007a).<br />
In sum, poverty is defined through the income-b<strong>as</strong>ed poverty<br />
guidelines issued by the federal government <strong>an</strong>d is widely used in<br />
research are<strong>as</strong> such <strong>as</strong> public health. <strong>SES</strong>, on the other h<strong>an</strong>d, is<br />
widely used in education research <strong>an</strong>d is generally defined more<br />
broadly th<strong>an</strong> poverty.<br />
May (2002) noted that good <strong>SES</strong> me<strong>as</strong>ures are needed in education<br />
because without them import<strong>an</strong>t inequalities may be overlooked<br />
or e<strong>as</strong>ily dismissed. However, there is signific<strong>an</strong>t concern<br />
over the quality of some of these me<strong>as</strong>ures, including the FRL<br />
variable (e.g., Entwisle & Astone, 1994; Hauser, 1994; Kurki<br />
et al., 2005). Central to these concerns is that <strong>SES</strong> itself is often<br />
poorly conceptualized by researchers, which affects the me<strong>as</strong>urement<br />
of <strong>SES</strong> <strong>an</strong>d the interpretation of statistical results (Haug,<br />
1977; Hauser & Warren, 1997; Oakes & Rossi, 2003).<br />
The Conceptualization of <strong>SES</strong> in Education Research<br />
Payne <strong>an</strong>d Biddle (1999) stated that research studying the effects<br />
of educational funding on achievement rarely defines <strong>SES</strong>, its<br />
theoretical roots, or its relationship to achievement. Un<strong>for</strong>tunately,<br />
education research involving <strong>SES</strong> is generally subject to Payne<br />
<strong>an</strong>d Biddle’s criticism. In most education research studies, <strong>SES</strong><br />
seems to be defined solely by the variable regarded <strong>as</strong> capturing<br />
this in<strong>for</strong>mation (e.g., FRL); that is, <strong>SES</strong> is whatever the me<strong>as</strong>ure<br />
of <strong>SES</strong> me<strong>as</strong>ures. None of the educational studies or reports cited<br />
earlier provided a conceptualization of <strong>SES</strong> that explained what<br />
it represented or its expected relationship to educational outcomes<br />
of interest, raising questions about what me<strong>as</strong>ures of <strong>SES</strong><br />
in these studies represent.<br />
For example, a report <strong>an</strong>alyzing charter school data from the<br />
National Assessment of Educational Progress (NAEP; NCES,<br />
2006a) provided a figure in which eligibility <strong>for</strong> <strong>an</strong> FRL appeared<br />
<strong>as</strong> a student-level variable, but the report offered no definition of<br />
<strong>SES</strong> or why the FRL variable w<strong>as</strong> <strong>an</strong> appropriate me<strong>as</strong>ure of <strong>SES</strong>.<br />
The use of FRL status <strong>as</strong> <strong>an</strong> <strong>SES</strong> me<strong>as</strong>ure also regularly appears<br />
in the Nation’s Report Card summaries (e.g., NCES, 2007,<br />
2008a). Studies by Borm<strong>an</strong> <strong>an</strong>d Hewes (2002), Lubienski <strong>an</strong>d<br />
Lubienski (2006), Molnar et al. (1999), <strong>an</strong>d Stein et al. (2008)<br />
provide additional examples of this practice. Payne <strong>an</strong>d Biddle’s<br />
(1999) criticism also applies to educational studies using <strong>SES</strong><br />
scales (e.g., Martin et al., 2004; NCES, 2004, 2006b, 2008c).<br />
The lack of conceptualization is problematic <strong>for</strong> several re<strong>as</strong>ons.<br />
What does <strong>SES</strong> represent in these studies? Differential<br />
access to economic resources, to social resources in the community,<br />
to social power (Rose & Pevalin, 2003), to desired education<br />
resources, or something else? If <strong>SES</strong> is treated <strong>as</strong> a construct, is it<br />
one dimensional (e.g., access to economic resources), two<br />
maRch 2010 121
dimensional (e.g., access to economic <strong>an</strong>d social resources), or<br />
something more? What is its expected relationship to various<br />
educational outcomes?<br />
A more desirable approach is to conceptualize <strong>SES</strong> by drawing<br />
on theories of <strong>SES</strong> (social cl<strong>as</strong>s, social stratification), such <strong>as</strong> those<br />
attributed to Max Weber (Breen, 2005) <strong>an</strong>d Emile Durkheim<br />
(Grusky, 2005). The <strong>SES</strong> literature devoted to theoretical models<br />
like those of Weber <strong>an</strong>d Durkheim <strong>an</strong>d methodologies <strong>for</strong> me<strong>as</strong>uring<br />
<strong>SES</strong> is v<strong>as</strong>t, <strong>an</strong>d good resources include the Encyclopedia of<br />
Social Theory (Harrington, Marshall, & Muller, 2006), works by<br />
Nam <strong>an</strong>d Powers (1983) <strong>an</strong>d Wright (2002), the University of<br />
Amsterdam website (http://www.sociosite.net/), <strong>an</strong>d the Report<br />
of the APA T<strong>as</strong>k Force on Socioeconomic Status (APA, 2007b).<br />
Un<strong>for</strong>tunately, education researchers who turn to this literature<br />
<strong>for</strong> guid<strong>an</strong>ce in conceptualizing <strong>SES</strong> will find signific<strong>an</strong>t<br />
disagreement on key definitions, terms, <strong>an</strong>d theoretical <strong>as</strong>sumptions<br />
(APA, 2007b; Campbell, 1983; Haug, 1977; Oakes &<br />
Rossi, 2003). For example, there is no widely agreed-upon definition<br />
of <strong>SES</strong> or consensus on whether <strong>SES</strong> consists of a relatively<br />
modest number of discrete cl<strong>as</strong>ses or is best represented <strong>as</strong> a semicontinuous<br />
variable with possibly hundreds of cl<strong>as</strong>ses.<br />
The Me<strong>as</strong>urement of <strong>SES</strong> in Education Research<br />
The me<strong>as</strong>urement of <strong>SES</strong> h<strong>as</strong> generated a signific<strong>an</strong>t literature<br />
devoted to developing me<strong>as</strong>ures consistent with particular theories,<br />
minimizing the cost <strong>an</strong>d intrusiveness of these me<strong>as</strong>ures, <strong>an</strong>d<br />
improving their psychometric properties (Cirino et al., 2002;<br />
Haug, 1977; Hauser & Warren, 1997; Nam & Powers, 1983;<br />
Oakes & Rossi, 2003). However, the lack of agreement in the<br />
<strong>SES</strong> literature on key definitions, terms, <strong>an</strong>d theoretical <strong>as</strong>sumptions<br />
h<strong>as</strong> import<strong>an</strong>t ramifications <strong>for</strong> me<strong>as</strong>uring student <strong>SES</strong> in<br />
education research.<br />
In the best possible c<strong>as</strong>e, researchers carefully describe what<br />
<strong>SES</strong> is intended to represent in their study, <strong>for</strong> example, a student’s<br />
access to economic <strong>an</strong>d social resources, <strong>an</strong>d then turn to<br />
the <strong>SES</strong> literature <strong>an</strong>d identify a theoretical model with clearly<br />
delineated me<strong>as</strong>ures consistent with their conceptualization of<br />
<strong>SES</strong>. This process would help to resolve import<strong>an</strong>t questions<br />
about whether the me<strong>as</strong>urement of student <strong>SES</strong> should be b<strong>as</strong>ed<br />
on a construct that is unidimensional or multidimensional<br />
(Bollen, Gl<strong>an</strong>ville, & Stecklov, 2001) <strong>an</strong>d whether <strong>SES</strong> me<strong>as</strong>ures<br />
should be a single item (variable) or a scale consisting of multiple<br />
items (variables; Hauser & Warren, 1997).<br />
However, even if researchers clearly conceptualize <strong>SES</strong>, they<br />
will encounter multiple theoretical models in the <strong>SES</strong> literature<br />
that vary in their definitions of <strong>SES</strong>, underlying theoretical<br />
<strong>as</strong>sumptions, <strong>an</strong>d appropriate me<strong>as</strong>ures <strong>an</strong>d that vary in their<br />
consistency with a researcher’s conceptualization. This leaves<br />
education researchers in the unhappy position of having to<br />
choose among competing (possibly unsatisfactory) models or to<br />
adapt a particular model <strong>an</strong>d use existing idiosyncratic me<strong>as</strong>ures<br />
(or develop their own) to fit their needs, <strong>an</strong>d then argue that this<br />
practice produces credible inferences (Oakes & Rossi, 2003).<br />
When researchers do not carefully conceptualize <strong>SES</strong>, turning<br />
to the theoretical <strong>SES</strong> literature <strong>for</strong> guid<strong>an</strong>ce is unlikely to be<br />
productive.<br />
What is needed is work connecting models of <strong>SES</strong> <strong>an</strong>d strategies<br />
<strong>for</strong> me<strong>as</strong>uring <strong>SES</strong> in educational contexts. A few examples<br />
122<br />
educational ReseaRcheR<br />
of this kind of connection have appeared, including work by<br />
Entwisle <strong>an</strong>d Astone (1994) in education <strong>an</strong>d Oakes <strong>an</strong>d Rossi<br />
(2003) in public health. An ef<strong>for</strong>t to generate better me<strong>as</strong>ures of<br />
<strong>SES</strong> is also under way at NCES (2008b), but the extent to which<br />
this work is in<strong>for</strong>med by theoretical models of <strong>SES</strong> is unclear.<br />
The <strong>Free</strong> <strong>Lunch</strong> Variable<br />
<strong>Eligibility</strong> <strong>for</strong> a <strong>Free</strong> or Reduced Price <strong>Lunch</strong><br />
<strong>Student</strong>s are certified <strong>as</strong> eligible <strong>for</strong> <strong>an</strong> FRL in one of two ways.<br />
One way relies on income in<strong>for</strong>mation provided by a householder.<br />
<strong>Student</strong>s are eligible <strong>for</strong> a reduced price lunch if their<br />
household income is less th<strong>an</strong> 185% of the federal poverty guidelines<br />
<strong>an</strong>d <strong>for</strong> a free lunch if their household income is less th<strong>an</strong><br />
130% of the poverty guidelines. For 2008, the official poverty<br />
guidelines <strong>for</strong> a family of four <strong>for</strong> the 48 contiguous states w<strong>as</strong> <strong>an</strong><br />
<strong>an</strong>nual income of $21,200, with Al<strong>as</strong>ka <strong>an</strong>d Hawaii having separate,<br />
<strong>an</strong>d higher, cutoffs (“Annual Update,” 2008).<br />
Using the poverty guidelines <strong>for</strong> 2008 <strong>for</strong> the 48 contiguous<br />
states, students living in a household of four whose income is less<br />
th<strong>an</strong> 1.85 × $21,200 = $39,220 would be certified <strong>as</strong> eligible <strong>for</strong><br />
a reduced price lunch, where<strong>as</strong> students from households whose<br />
income is less th<strong>an</strong> 1.3 × $21,200 = $27,560 would be certified<br />
<strong>as</strong> eligible <strong>for</strong> a free lunch. Data available <strong>for</strong> the 2007–2008<br />
school year indicate that 92% of all K–12 students in the United<br />
States had access to <strong>an</strong> FRL, which is less th<strong>an</strong> 100% because<br />
school district participation is voluntary. At the end of the 2006–<br />
2007 school year, approximately 18.4 million children received<br />
<strong>an</strong> FRL, or about 60% of all school lunches served (Food <strong>an</strong>d<br />
Nutrition Service [FNS], 2008).<br />
A second avenue to eligibility is direct certification, b<strong>as</strong>ed on<br />
whether a household receives food stamps, h<strong>as</strong> foster children in<br />
the home, or participates in at le<strong>as</strong>t one federally funded <strong>as</strong>sist<strong>an</strong>ce<br />
program such <strong>as</strong> WIC or TANF (FNS, 2008). <strong>Student</strong>s in<br />
a direct certification household are not generally required to submit<br />
<strong>an</strong> application <strong>for</strong> free meals; rather, social service agencies<br />
work closely with schools to identify students eligible <strong>for</strong> direct<br />
certification.<br />
Prominence of the <strong>Free</strong> <strong>Lunch</strong><br />
Variable in Education Research<br />
Nierm<strong>an</strong> <strong>an</strong>d Veak (1997) argued that the eligibility of a child <strong>for</strong><br />
<strong>an</strong> FRL b<strong>as</strong>ed on family income <strong>an</strong>d its relation to the poverty<br />
line have led education researchers to use this variable <strong>as</strong> a me<strong>as</strong>ure<br />
of <strong>SES</strong>. Kurki et al. (2005) claimed that FRL is the most<br />
commonly used me<strong>as</strong>ure of poverty in education research but<br />
offered no evidence supporting that <strong>as</strong>sertion. A review of studies<br />
of educational achievement by Harwell, Maeda, <strong>an</strong>d Lee (2004)<br />
found that during the period from 1996 to 2004 approximately<br />
20% of the articles published in the Americ<strong>an</strong> Educational<br />
Research Journal <strong>an</strong>d Sociology of Education employing <strong>an</strong> <strong>SES</strong><br />
me<strong>as</strong>ure used the FRL variable. Sirin’s (2005) meta-<strong>an</strong>alysis of<br />
the relationship between <strong>SES</strong> <strong>an</strong>d achievement <strong>for</strong> studies published<br />
between 1990 <strong>an</strong>d 2000 reported that approximately 17%<br />
of the sampled studies used FRL <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong>.<br />
It is apparent that education researchers continue to use the<br />
FRL variable <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong> (e.g., Alex<strong>an</strong>der, Entwisle, &<br />
Olson, 2007; Liew, McTigue, Barrois, & Hughes, 2008;
Lubienski & Lubienski, 2006; Stein et al., 2008). FRL also continues<br />
to be used in federally funded studies <strong>an</strong>d reports with<br />
import<strong>an</strong>t policy implications (e.g., Martin et al., 2004; NCES,<br />
2006a, 2007, 2008a) <strong>an</strong>d studies connected to federally funded<br />
education research centers (e.g., Boykin, Colem<strong>an</strong>, Lilja, & Tyler,<br />
2004; W<strong>an</strong>g, Niemi, & W<strong>an</strong>g, 2007). The fact that the FRL<br />
variable w<strong>as</strong> described <strong>as</strong> <strong>an</strong> indicator of economic adv<strong>an</strong>tage<br />
in the No Child Left Behind (2002) legislation, <strong>an</strong>d continues<br />
to be described in this way (Spellings, 2007), may also encourage<br />
its use.<br />
The NSLP <strong>an</strong>d Its Current Operation in U.S. Schools<br />
The origins of offering free <strong>an</strong>d reduced price lunches c<strong>an</strong> be<br />
traced to early Europe<strong>an</strong> <strong>an</strong>d U.S. programs designed to feed<br />
hungry children (Gunderson, 2003). However, the impetus <strong>for</strong><br />
large-scale federal involvement came in response to evidence that<br />
men from poor families were disproportionately denied admitt<strong>an</strong>ce<br />
to the armed services during World War II because of physical<br />
problems <strong>as</strong>sociated with poor nutrition (Dev<strong>an</strong>ey, Ellwood,<br />
& Love, 1997). This provided the impetus <strong>for</strong> the Richard B.<br />
Russell National School <strong>Lunch</strong> Act (NSLA), which w<strong>as</strong> signed<br />
into law by President Harry Trum<strong>an</strong> in 1946. The goal of the<br />
NSLA w<strong>as</strong> to promote the health <strong>an</strong>d well-being of children <strong>an</strong>d<br />
incre<strong>as</strong>e student learning by providing a low-cost healthy meal.<br />
The NSLP is part of the NSLA (Economic Research Service<br />
[ERS], 2008).<br />
This program makes af<strong>for</strong>dable <strong>an</strong>d healthy lunches available<br />
to all K–12 students, m<strong>an</strong>y free or at a reduced price, <strong>an</strong>d operates<br />
in both public <strong>an</strong>d nonprofit private schools. Every student<br />
who receives a school lunch h<strong>as</strong> some portion of the cost covered<br />
by this program (NSLP, 2008). Levine (2008) provides <strong>an</strong><br />
account of the NSLP that emph<strong>as</strong>izes the role of nutrition in the<br />
program’s origins <strong>an</strong>d goals.<br />
The legislative history of the NLSP (Gunderson, 2003) provides<br />
evidence of ch<strong>an</strong>ges in this program in U.S. schools, which<br />
are reflected in incre<strong>as</strong>es in student participation. Figure 1 shows<br />
that the average daily participation rate <strong>for</strong> FRLs h<strong>as</strong> generally<br />
incre<strong>as</strong>ed since 1969, the largest incre<strong>as</strong>e being in the number of<br />
free lunches served. On the other h<strong>an</strong>d, the number of full-price<br />
lunches served h<strong>as</strong> generally decre<strong>as</strong>ed. Figure 1 also shows a drop<br />
in the total number of lunches served around 1980—which corresponds<br />
to a drop in the upper limit <strong>for</strong> reduced price lunches<br />
from 195% to 185% of the poverty line (Gunderson, 2003)—<br />
but <strong>an</strong> upward trend thereafter.<br />
Today the NSLP is deeply ingrained in the U.S. educational<br />
system, operating in more th<strong>an</strong> 100,000 public<br />
schools, nonprofit private schools, <strong>an</strong>d state-licensed facilities<br />
that provide residential child care services (Levine, 2008;<br />
NSLP, 2008).<br />
Figure 2 summarizes the general operation of the NSLP <strong>an</strong>d<br />
provides ample evidence of its complex, multilayered structure.<br />
The program begins with the U.S. Department of Agriculture,<br />
which works with states (usually state departments of education),<br />
which in turn work with local education agencies (LEAs), which<br />
could be a school board, <strong>an</strong> administrative agency, or a combination<br />
of school districts or counties recognized by a state. Among<br />
the m<strong>an</strong>y t<strong>as</strong>ks of LEAs <strong>an</strong>d schools are ensuring that households<br />
are aware of the program, determining student eligibility <strong>for</strong> <strong>an</strong><br />
Number of <strong>Student</strong>s (in Millions)<br />
0 5 10 15 20 25 30 35<br />
Total<br />
<strong>Free</strong><br />
Reduced Price<br />
Full Price<br />
1970 1975 1980 1985 1990<br />
Year<br />
1995 2000 2005<br />
FIGURE 1. National School <strong>Lunch</strong> Program participation rates<br />
over time.<br />
FIGURE 2. Outline of the National School <strong>Lunch</strong> Program.<br />
USDA = U.S. Department of Agriculture.<br />
FRL, implementing the NSLP in the schools, <strong>an</strong>d keeping<br />
records needed by the state <strong>an</strong>d the FNS. Descriptions of the<br />
components of the NSLP appear in Gunderson (2003), Levine<br />
(2008), <strong>an</strong>d on the NSLP (2008) website.<br />
The current operation of the NSLP in schools makes it clear<br />
that this is a large, complex program with m<strong>an</strong>y facets; that student<br />
eligibility <strong>for</strong> <strong>an</strong> FRL is not strictly income b<strong>as</strong>ed; <strong>an</strong>d that<br />
the core purpose of the program continues to be to incre<strong>as</strong>e<br />
maRch 2010 123
student nutrition <strong>an</strong>d learning. These characteristics provide a<br />
context <strong>for</strong> examining the FRL variable.<br />
FRL Particip<strong>an</strong>ts<br />
Surprisingly little is known about the demographics of FRL particip<strong>an</strong>ts<br />
(ERS, 2006). Recently, national data sets such <strong>as</strong> the<br />
Survey of Income <strong>an</strong>d Program Participation, the National<br />
Health <strong>an</strong>d Nutrition Examination Survey, <strong>an</strong>d the Public<br />
Elementary/Secondary School Universe Survey (NCES, 2003–<br />
2004) have been used to generate profiles of FRL particip<strong>an</strong>ts <strong>for</strong><br />
the purpose of better in<strong>for</strong>ming researchers <strong>an</strong>d program administrators.<br />
An ERS (2006) report b<strong>as</strong>ed on Survey of Income <strong>an</strong>d<br />
Program Participation data indicated that participation rates<br />
were highest <strong>for</strong> elementary school–age students <strong>an</strong>d declined<br />
each year thereafter. The participation of White, Afric<strong>an</strong><br />
Americ<strong>an</strong>, <strong>an</strong>d Hisp<strong>an</strong>ic students in the FRL program w<strong>as</strong> proportionally<br />
equal, although White students, compared with<br />
Afric<strong>an</strong> Americ<strong>an</strong> <strong>an</strong>d Hisp<strong>an</strong>ic students, were more likely to<br />
receive a reduced price lunch th<strong>an</strong> a free lunch.<br />
The results in Table 1 suggest that schools with more Afric<strong>an</strong><br />
Americ<strong>an</strong> students tend to have more students eligible <strong>for</strong> <strong>an</strong><br />
FRL. For example, among schools with 75% or more Afric<strong>an</strong><br />
Americ<strong>an</strong> students, 32.2% were eligible <strong>for</strong> <strong>an</strong> FRL, where<strong>as</strong> <strong>for</strong><br />
schools with 75% or more White students this percentage w<strong>as</strong><br />
3.6%.<br />
Table 2 shows that the location of a school appears to be<br />
related to how m<strong>an</strong>y students are certified <strong>as</strong> eligible <strong>for</strong> <strong>an</strong> FRL<br />
within that school. Schools in urb<strong>an</strong> are<strong>as</strong> tend to have more<br />
students eligible <strong>for</strong> <strong>an</strong> FRL, <strong>an</strong>d students in schools in a suburb,<br />
smaller town, or rural area tend to have fewer students eligible.<br />
Characteristics of the <strong>Free</strong> <strong>Lunch</strong> Variable<br />
<strong>Student</strong> eligibility <strong>for</strong> <strong>an</strong> FRL h<strong>as</strong> several characteristics that are<br />
import<strong>an</strong>t in underst<strong>an</strong>ding its widespread use <strong>an</strong>d its deficiencies<br />
<strong>as</strong> <strong>an</strong> <strong>SES</strong> me<strong>as</strong>ure.<br />
<strong>Eligibility</strong> <strong>for</strong> <strong>an</strong> FRL is a poor me<strong>as</strong>ure of a student’s access to economic<br />
resources. <strong>Student</strong> eligibility <strong>for</strong> <strong>an</strong> FRL is not strictly b<strong>as</strong>ed<br />
on federal poverty guidelines. As described above, students in a<br />
household of four are eligible <strong>for</strong> a free lunch if household income<br />
is less th<strong>an</strong> $21,200 × 1.3 = $27,560. Correspondingly, students<br />
are eligible <strong>for</strong> a reduced price lunch if household income is less<br />
124<br />
educational ReseaRcheR<br />
Table 1<br />
Percentage of <strong>Student</strong>s Eligible <strong>for</strong> <strong>Free</strong> <strong>an</strong>d Reduced Price <strong>Lunch</strong>es by Race/Ethnicity<br />
<strong>for</strong> a Sample of Public Elementary <strong>an</strong>d Secondary Schools <strong>for</strong> 2003–2004<br />
Race/Ethnicity<br />
10% or Less 11%–25% 26%–50% 51%–75% 75% or More<br />
Total 15.00 20.60 29.10 20.30 15.00<br />
White 20.70 27.60 33.20 14.90 3.60<br />
Black 4.20 9.40 24.50 29.80 32.20<br />
Hisp<strong>an</strong>ic 6.00 9.20 21.70 28.40 34.70<br />
Asi<strong>an</strong>/Pacific Isl<strong>an</strong>der 22.50 22.00 26.40 18.20 10.90<br />
Americ<strong>an</strong> Indi<strong>an</strong>/Al<strong>as</strong>ka Native 5.10 12.20 27.50 29.80 25.50<br />
th<strong>an</strong> $21,200 × 1.85 = $39,220. Thus, eligibility <strong>for</strong> <strong>an</strong> FRL<br />
depends indirectly on federal poverty guidelines.<br />
Hauser (1994) argued convincingly that the federal poverty<br />
guidelines are flawed <strong>an</strong>d outdated, <strong>an</strong>d hence variables such <strong>as</strong><br />
FRL that are b<strong>as</strong>ed on the guidelines are similarly flawed. Hauser<br />
also identified several problems in using the poverty guidelines <strong>as</strong><br />
the b<strong>as</strong>is of <strong>an</strong> <strong>SES</strong> me<strong>as</strong>ure.<br />
M<strong>an</strong>y of these problems are economic in nature <strong>an</strong>d center on<br />
the failure of the guidelines to take in-kind benefits into account<br />
in determining eligibility in a way that considers the fin<strong>an</strong>cial<br />
impact of these benefits, rather th<strong>an</strong> simply triggering eligibility.<br />
Among the largest <strong>an</strong>d best known in-kind benefits programs are<br />
WIC, the Supplemental Nutrition Assist<strong>an</strong>ce Program (food<br />
stamps), housing <strong>as</strong>sist<strong>an</strong>ce, <strong>an</strong>d Head Start. An incre<strong>as</strong>ingly<br />
import<strong>an</strong>t in-kind benefit is the Earned Income Tax Credit. In<br />
2008, a low-income family with one child could receive a direct<br />
income tax credit up to 8% of its income, <strong>an</strong>d a low-income family<br />
with two or more children could receive a credit of up to<br />
11.5% (Internal Revenue Service, 2008).<br />
Hauser also noted that there are other economic factors not<br />
taken into account in the federal poverty guidelines, including the<br />
cost of earning income, child care costs, geographic variation in<br />
the cost of living, direct tax payments such <strong>as</strong> payroll <strong>an</strong>d income<br />
taxes, differences in health insur<strong>an</strong>ce coverage, <strong>an</strong>d the fact that<br />
poverty guidelines have not been updated since they were created<br />
in the 1960s to account <strong>for</strong> ch<strong>an</strong>ging consumption patterns.<br />
Kurki et al. (2005) raised similar concerns about the FRL variable.<br />
These authors emph<strong>as</strong>ized that the FRL variable does not<br />
necessarily capture relev<strong>an</strong>t dimensions of poverty, <strong>for</strong> example,<br />
the effects of concentrated poverty in a neighborhood.<br />
<strong>Student</strong>s c<strong>an</strong> also be directly certified <strong>as</strong> eligible <strong>for</strong> <strong>an</strong> FRL in<br />
ways that do not depend directly on household income. As noted<br />
earlier, students living in foster homes or in households that participate<br />
in federal <strong>as</strong>sist<strong>an</strong>ce programs like WIC are automatically<br />
certified <strong>as</strong> eligible, <strong>an</strong>d approximately 25% of students are certified<br />
<strong>as</strong> eligible in this way (ERS, 2003). In these c<strong>as</strong>es, there is no<br />
requirement that household income be less th<strong>an</strong> 130% or 185%<br />
of the poverty guidelines <strong>for</strong> a student to be eligible <strong>for</strong> <strong>an</strong> FRL.<br />
Compounding these difficulties is compelling evidence that<br />
a signific<strong>an</strong>t percentage of students are incorrectly certified <strong>as</strong><br />
eligible or not eligible. Hauser (1994) <strong>an</strong>d Gle<strong>as</strong>on (2008) commented<br />
that the determination of student eligibility <strong>for</strong> <strong>an</strong> FRL<br />
traditionally h<strong>as</strong> placed relatively little emph<strong>as</strong>is on verification
Table 2<br />
Percentage of Public Elementary Schools <strong>an</strong>d Secondary <strong>Student</strong>s by Percentage of <strong>Student</strong>s<br />
Eligible <strong>for</strong> <strong>Free</strong> <strong>an</strong>d Reduced Price <strong>Lunch</strong>es <strong>an</strong>d by School Location <strong>for</strong> 2003–2004<br />
Location 10% or Less 11%–25% 26%–50% 51%–75% 75% or More<br />
City 8.60 13.20 24.70 24.50 29.10<br />
Suburb 25.40 25.90 25.00 14.50 9.20<br />
Town 6.00 19.20 39.10 25.20 10.50<br />
Rural area 11.70 22.60 36.20 21.40 8.10<br />
of the eligibility of applic<strong>an</strong>ts, <strong>an</strong>d there is subst<strong>an</strong>tial evidence<br />
supporting this <strong>as</strong>sertion.<br />
An income verification study using a nationally representative<br />
sample showed that 17% of students certified <strong>as</strong> eligible <strong>for</strong> <strong>an</strong><br />
FRL should not have been, <strong>an</strong>d 8% certified <strong>as</strong> ineligible were in<br />
fact eligible (FNS, 1990). Another study (Office of Inspector<br />
General, 1997) used data from Illinois <strong>an</strong>d showed that 19% of<br />
students certified <strong>as</strong> eligible should not have been, <strong>an</strong>d a 2003<br />
audit by the General Accounting Office estimated that 21% of<br />
the sampled households were incorrectly certified <strong>as</strong> eligible or<br />
not eligible <strong>for</strong> <strong>an</strong> FRL. A study of more th<strong>an</strong> 3,600 students in<br />
21 school districts reported that 20% of the students certified <strong>as</strong><br />
eligible <strong>for</strong> <strong>an</strong> FRL should not have been, <strong>an</strong>d 7% certified <strong>as</strong> not<br />
eligible were eligible (Office of Research, Nutrition, <strong>an</strong>d Analysis<br />
[ORNA], 2004). A 2005 study found that 6% to 9% of students<br />
who were declared eligible using direct certification should not<br />
have been (Mathematica Policy Research, 2005).<br />
The most recent large certification study involved 356 schools<br />
<strong>an</strong>d more th<strong>an</strong> 8,000 students. The results indicated that 15.8%<br />
of the students certified <strong>as</strong> eligible <strong>for</strong> a free lunch should not<br />
have been <strong>an</strong>d 6% certified <strong>as</strong> eligible <strong>for</strong> a reduced price lunch<br />
were actually eligible <strong>for</strong> a free lunch (ORNA, 2007), <strong>for</strong> a total<br />
certification error rate of 21.8%. Incorrectly certifying students<br />
<strong>as</strong> eligible appears to be more common th<strong>an</strong> incorrectly certifying<br />
them <strong>as</strong> not eligible (ORNA, 2007). This study also found that<br />
most certification error originates in the eligibility in<strong>for</strong>mation<br />
provided by households rather th<strong>an</strong> in administrative error. For<br />
example, 23% of the households in the ORNA study incorrectly<br />
reported income or household size on the FRL application.<br />
Another source of error occurs when students are certified <strong>as</strong><br />
eligible or not eligible in <strong>an</strong> ad hoc way. A recent ERS (2008)<br />
study found that students who failed to return <strong>an</strong> FRL application<br />
or whose application w<strong>as</strong> incomplete were typically categorized<br />
by local officials (e.g., the local education authority or its<br />
agent, or the school director of food services) <strong>as</strong> not eligible. In<br />
<strong>an</strong> <strong>an</strong>alysis of NAEP charter school data (NCES, 2006a), students<br />
<strong>for</strong> whom no FRL in<strong>for</strong>mation w<strong>as</strong> obtained because<br />
school records were not available or the school did not offer FRLs<br />
were simply cl<strong>as</strong>sified <strong>as</strong> not eligible. Lubienski <strong>an</strong>d Lubienski<br />
(2006) also <strong>an</strong>alyzed the NAEP charter school data <strong>an</strong>d imputed<br />
a student’s FRL status on the b<strong>as</strong>is of either the reported percentage<br />
of eligible students at a school or whether a student scored<br />
above a cutoff on a home resources checklist. The fact that<br />
the fin<strong>an</strong>cial cost of providing <strong>an</strong> FRL is low relative to the cost<br />
<strong>as</strong>sociated with m<strong>an</strong>y other federal programs also contributes to<br />
ad hoc certification error because the cost of incorrectly certifying<br />
a student <strong>as</strong> eligible is small (Hauser, 1994; ORNA, 2007).<br />
It is also possible that certification error is incre<strong>as</strong>ing in the No<br />
Child Left Behind era. <strong>Student</strong>s deliberately miscl<strong>as</strong>sified <strong>as</strong> eligible<br />
may help to raise average test scores, incre<strong>as</strong>ing the likelihood<br />
that a school or a key student subgroup will be cl<strong>as</strong>sified <strong>as</strong><br />
making adequate yearly progress.<br />
Figure 1 shows a clear incre<strong>as</strong>e in the percentage of students<br />
eligible <strong>for</strong> a free lunch starting in 2001–2002. Although not<br />
conclusive, this incre<strong>as</strong>e is consistent with the introduction of No<br />
Child Left Behind, which w<strong>as</strong> signed into law in J<strong>an</strong>uary 2002,<br />
<strong>an</strong>d with evidence that schools sometimes deliberately exclude<br />
groups of students from schoolwide testing in the hope of incre<strong>as</strong>ing<br />
test scores (Nichols & Berliner, 2008).<br />
In short, available evidence suggests that researchers using<br />
FRL <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong> c<strong>an</strong> expect a signific<strong>an</strong>t percentage of<br />
students, perhaps <strong>as</strong> high <strong>as</strong> 20%, to be miscl<strong>as</strong>sified. The effect<br />
of miscl<strong>as</strong>sification is that the magnitude <strong>an</strong>d nature of the relationship<br />
between <strong>an</strong> outcome variable <strong>an</strong>d the FRL variable c<strong>an</strong><br />
be distorted, producing bi<strong>as</strong>ed inferences.<br />
For example, Stein et al. (2008) used FRL <strong>as</strong> a student-level<br />
covariate in a multilevel model of reading achievement gains in<br />
which 62% of the sample w<strong>as</strong> eligible <strong>for</strong> <strong>an</strong> FRL. The resulting<br />
fixed effects slope <strong>for</strong> FRL of −1.41 <strong>for</strong> Model 1 w<strong>as</strong> statistically<br />
signific<strong>an</strong>t, me<strong>an</strong>ing that, with other covariates held const<strong>an</strong>t,<br />
students eligible <strong>for</strong> <strong>an</strong> FRL were expected to show a decline from<br />
pretest to posttest in reading achievement. It is not possible to<br />
know <strong>for</strong> certain what the impact of a 20% miscl<strong>as</strong>sification rate<br />
would be, but it is likely that the slope of −1.41 would be too<br />
large or too small (i.e., is bi<strong>as</strong>ed). For example, suppose most<br />
miscl<strong>as</strong>sified students were cl<strong>as</strong>sified <strong>as</strong> eligible even though they<br />
came from households with greater access to economic resources.<br />
Plausible outcomes of this miscl<strong>as</strong>sification are smaller differences<br />
between those eligible <strong>an</strong>d not eligible <strong>an</strong>d a downward<br />
bi<strong>as</strong>ing of the slope of −1.41 (i.e., making the slope too small).<br />
Difficulties linked to the use of FRL <strong>as</strong> a me<strong>as</strong>ure of a student’s<br />
access to economic resources are aggravated by me<strong>as</strong>urement<br />
issues related to its dichotomous (eligible, not eligible) scaling.<br />
For example, using the 2008 poverty guidelines, a student from<br />
a household of four with <strong>an</strong> income of $150,000 h<strong>as</strong> the same<br />
FRL status <strong>as</strong> a student from a household of four with <strong>an</strong> income<br />
of $50,000. Such a large difference in household income implies<br />
potentially import<strong>an</strong>t differences in access to economic resources<br />
that the FRL variable c<strong>an</strong>not capture. Along the same lines, a<br />
student from a household of four with <strong>an</strong> income of $40,000 <strong>an</strong>d<br />
a student from a household of four with <strong>an</strong> income of $38,000<br />
have a different FRL status. Such a modest difference in household<br />
income implies that students have similar access to economic<br />
resources yet have a different <strong>SES</strong> status.<br />
maRch 2010 125
Participation rates decline with incre<strong>as</strong>es in grade level. An import<strong>an</strong>t<br />
characteristic of the FRL variable is the decline in participation<br />
rates with incre<strong>as</strong>es in grade level. In general, the highest<br />
participation rates occur <strong>for</strong> elementary school students <strong>an</strong>d then<br />
decline subst<strong>an</strong>tially <strong>for</strong> middle school <strong>an</strong>d high school students.<br />
A report by the ERS (2006) provided evidence that FRL participation<br />
w<strong>as</strong> greatest <strong>for</strong> the 8 to 13 age r<strong>an</strong>ge <strong>an</strong>d lowest <strong>for</strong> the<br />
16 to 18 age r<strong>an</strong>ge; <strong>an</strong>other report found that FRL participation<br />
peaked <strong>for</strong> students 8 to 10 years old (32.8%) <strong>an</strong>d declined subst<strong>an</strong>tially<br />
among 14- to 15-year-olds (12.5%) <strong>an</strong>d 16- to 18-yearolds<br />
(13.5%; ERS, 2008). The decline in participation w<strong>as</strong><br />
overwhelmingly due to the failure of students to return application<br />
<strong>for</strong>ms <strong>for</strong> FRL eligibility. There is evidence of a stigma <strong>as</strong>sociated<br />
with receiving <strong>an</strong> FRL among parents <strong>an</strong>d middle school<br />
<strong>an</strong>d high school students that may play <strong>an</strong> import<strong>an</strong>t role in the<br />
decline in participation (ORNA, 1994), but this h<strong>as</strong> not been<br />
carefully studied.<br />
The lower rates of student participation at all levels of K–12,<br />
but especially in middle school <strong>an</strong>d high school, raise import<strong>an</strong>t<br />
questions about the use of FRL eligibility <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong>. A<br />
related concern is the apparently common practice of certifying<br />
students who fail to return applications <strong>for</strong> FRL eligibility <strong>as</strong> not<br />
eligible (ERS, 2008). In this c<strong>as</strong>e, inferences from <strong>an</strong>alyses involving<br />
the FRL variable would likely be bi<strong>as</strong>ed.<br />
FRL status is available <strong>for</strong> every student. FRL status is expected to<br />
be available <strong>for</strong> all students, which is import<strong>an</strong>t because of the<br />
potential impact of nonresponse (missing data) <strong>an</strong>d nonresponse<br />
bi<strong>as</strong>. Nonresponse occurs when students or households fail to<br />
provide requested in<strong>for</strong>mation, such <strong>as</strong> household income, <strong>an</strong>d<br />
nonresponse bi<strong>as</strong> occurs when data are missing systematically<br />
(e.g., wealthier households are less likely to provide income in<strong>for</strong>mation),<br />
truncating the sample <strong>an</strong>d bi<strong>as</strong>ing inferences (Groves,<br />
2006). However, schools certify students <strong>as</strong> eligible or not eligible<br />
<strong>for</strong> <strong>an</strong> FRL even if householder income in<strong>for</strong>mation is not available,<br />
so the FRL variable is not subject to nonresponse bi<strong>as</strong>.<br />
Moreover, FRL status is typically available in <strong>an</strong> electronic <strong>for</strong>m,<br />
which keeps costs down.<br />
There is confounding of the FRL effect. Hauser (1994) pointed out<br />
<strong>an</strong>other characteristic that is related to FRL eligibility: Comparing<br />
students eligible <strong>for</strong> <strong>an</strong> FRL <strong>an</strong>d those not eligible confounds the<br />
effects of low <strong>SES</strong> <strong>an</strong>d the presumably positive effects of the program<br />
itself. For example, the finding that eligible students score<br />
on average lower on a st<strong>an</strong>dardized mathematics test may be<br />
partly attributable to less access to economic <strong>an</strong>d social resources<br />
but may also be related to receiving a free lunch. That is, students<br />
might show even poorer achievement if they did not receive a free<br />
lunch.<br />
The above characteristics indicate that the FRL variable is<br />
consistently defined because of its link to the government’s poverty<br />
guidelines, is available <strong>for</strong> every student, is nonintrusive, is<br />
simple (eligible, not eligible), <strong>an</strong>d c<strong>an</strong> be obtained relatively<br />
cheaply because in<strong>for</strong>mation is taken directly from school<br />
records. These characteristics help to explain the continued<br />
popularity of this variable <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong>. On the other<br />
h<strong>an</strong>d, the facts that students are sometimes certified <strong>as</strong> eligible<br />
b<strong>as</strong>ed on nonincome factors, that eligibility status will likely<br />
126<br />
educational ReseaRcheR<br />
show signific<strong>an</strong>t miscl<strong>as</strong>sification error, <strong>an</strong>d that receiving <strong>an</strong><br />
FRL fundamentally represents <strong>an</strong> intervention raises concerns<br />
about the validity of inferences b<strong>as</strong>ed on this variable.<br />
FRL <strong>an</strong>d Other Me<strong>as</strong>ures of <strong>SES</strong><br />
Indications that the FRL variable is a poor me<strong>as</strong>ure of <strong>SES</strong> suggest<br />
that other variables should be considered. Identifying other<br />
me<strong>as</strong>ures begins with a researcher providing a clear conceptualization<br />
of <strong>SES</strong> <strong>an</strong>d selecting one or more variables that capture<br />
the relev<strong>an</strong>t circumst<strong>an</strong>ces of a student’s <strong>SES</strong> consistent with that<br />
conceptualization <strong>an</strong>d with the purpose of the study at h<strong>an</strong>d.<br />
In m<strong>an</strong>y c<strong>as</strong>es, the conceptualization will be characterized by<br />
a student’s access to economic resources; in others, the intent may<br />
be to <strong>as</strong>sess a student’s access to social capital (contacts in networks<br />
that c<strong>an</strong> lead to personal or professional gains) or to identify<br />
variables that generate a hierarchical ordering of households<br />
in a community.<br />
Once a researcher h<strong>as</strong> described his or her conceptualization<br />
of <strong>SES</strong>, the criteria recommended by Hauser (1994), Entwisle<br />
<strong>an</strong>d Astone (1994), Oakes <strong>an</strong>d Rossi (2003), <strong>an</strong>d Hobbs <strong>an</strong>d<br />
Vignoles (2007) c<strong>an</strong> be used to identify appropriate <strong>SES</strong> me<strong>as</strong>ures.<br />
These criteria suggest the following <strong>for</strong> selected me<strong>as</strong>ures<br />
of <strong>SES</strong>:<br />
1. <strong>SES</strong> me<strong>as</strong>ures should validly <strong>an</strong>d reliably capture <strong>SES</strong> <strong>as</strong><br />
conceptualized by the researcher. For example, defining<br />
<strong>SES</strong> <strong>as</strong> a student’s access to economic resources should<br />
result in the use of variables known to accurately <strong>an</strong>d consistently<br />
capture this in<strong>for</strong>mation <strong>an</strong>d should provide data<br />
comparable to those from other sources, that is, should<br />
provide in<strong>for</strong>mation that is consistent across various<br />
sources. This provides evidence of convergent validity.<br />
2. <strong>SES</strong> me<strong>as</strong>ures should show participation rates that are<br />
unrelated to a student’s grade level.<br />
3. <strong>SES</strong> me<strong>as</strong>ures should show minimal nonresponse. This<br />
is import<strong>an</strong>t, <strong>as</strong> nonresponse bi<strong>as</strong> c<strong>an</strong> produce invalid<br />
inferences.<br />
4. <strong>SES</strong> me<strong>as</strong>ures should be accessible at a re<strong>as</strong>onable cost.<br />
Also, the selected me<strong>as</strong>ure should not confound the relationship<br />
between <strong>SES</strong> <strong>an</strong>d <strong>an</strong> educational outcome. Of course, it may not<br />
be possible <strong>for</strong> a single me<strong>as</strong>ure to satisfy all of these criteria<br />
(Hauser, 1994).<br />
Me<strong>as</strong>ures possessing m<strong>an</strong>y of the above characteristics have<br />
been used in various literatures <strong>for</strong> some time, including education.<br />
These include householder income, education, <strong>an</strong>d occupation<br />
<strong>an</strong>d variables intended to capture home resources. An impressive<br />
feature of these variables is the extensive research studying their<br />
use <strong>as</strong> me<strong>as</strong>ures of <strong>SES</strong> in various literatures (e.g., Cirino et al.,<br />
2002; Edwards-Hewitt & Gray, 1995; Hauser & Warren, 1997;<br />
Jones & McMill<strong>an</strong>, 2001; Nam & Powers, 1983; Rose & Pevalin,<br />
2003). This literature makes it clear that these me<strong>as</strong>ures have<br />
m<strong>an</strong>y strengths but also some weaknesses, particularly nonresponse<br />
by households unwilling to provide in<strong>for</strong>mation <strong>an</strong>d the<br />
cost of collecting the in<strong>for</strong>mation. These weaknesses have likely<br />
contributed to the ongoing use of FRL <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong>.
<strong>SES</strong> Me<strong>as</strong>ure<br />
<strong>Free</strong> or reduced<br />
price lunch<br />
eligibility<br />
Table 3<br />
Comparing Socioeconomic Status (<strong>SES</strong>) Me<strong>as</strong>ures Used in Education Research<br />
Valid Indicator of<br />
Access to Economic Resources<br />
Highly unlikely because of its<br />
failure to adequately capture<br />
household economic<br />
resources<br />
Householder income Likely, because it adequately<br />
captures household<br />
economic resources<br />
Householder<br />
occupation<br />
Likely because it adequately<br />
captures household<br />
economic resources<br />
Number of siblings Highly unlikely because it<br />
does not adequately capture<br />
household economic<br />
Home resources<br />
checklist<br />
resources<br />
Somewhat likely; depends<br />
heavily on breadth <strong>an</strong>d<br />
depth of the checklist items<br />
Table 3 is used to facilitate comparisons of several variables<br />
used <strong>as</strong> me<strong>as</strong>ures of <strong>SES</strong> in education research following the recommendations<br />
of Hauser (1994), Entwisle <strong>an</strong>d Astone (1994),<br />
Oakes <strong>an</strong>d Rossi (2003), <strong>an</strong>d Hobbs <strong>an</strong>d Vignoles (2007). In<br />
examining Table 3, it is import<strong>an</strong>t to realize that each application<br />
of <strong>SES</strong> me<strong>as</strong>ures in a study reflects conditions unique to that<br />
study <strong>an</strong>d that choosing a me<strong>as</strong>ure depends heavily on the purpose<br />
of the study (Entwisle & Astone, 1994). Accordingly, Table<br />
3 should be viewed <strong>as</strong> providing general guid<strong>an</strong>ce <strong>for</strong> comparing<br />
<strong>SES</strong> me<strong>as</strong>ures. More extensive guid<strong>an</strong>ce would incorporate a<br />
broad r<strong>an</strong>ge of factors believed to affect educational outcomes<br />
(e.g., cultural influence, concentrations of poverty) <strong>an</strong>d <strong>as</strong>sociated<br />
me<strong>as</strong>ures.<br />
There’s No <strong>Free</strong> <strong>Lunch</strong><br />
The comparisons in Table 3 suggest three general conclusions.<br />
First, when nonresponse is minimal (e.g., 1%–2%) <strong>an</strong>d access<br />
<strong>an</strong>d cost are not major concerns, householder income, education,<br />
<strong>an</strong>d occupation have much to recommend them. If these<br />
Participation<br />
Rates Const<strong>an</strong>t<br />
Across Grades<br />
Minimal<br />
Nonresponse<br />
No Yes, because schools<br />
certify each student <strong>as</strong><br />
eligible/not eligible<br />
Likely Somewhat likely to<br />
somewhat unlikely;<br />
depends on student/<br />
householder responses<br />
to a survey (e.g.,<br />
nonresponse rate of<br />
12%–25%)<br />
Likely Somewhat likely to<br />
somewhat unlikely;<br />
depends on student/<br />
householder responses to<br />
a survey (e.g., 6%–25%)<br />
Likely Yes Likely<br />
Likely Somewhat likely to<br />
somewhat unlikely;<br />
depends on student/<br />
householder responses to<br />
a survey (e.g., 1%–30%)<br />
Highly likely<br />
E<strong>as</strong>y Access,<br />
Inexpensive<br />
Somewhat unlikely;<br />
depends on student/<br />
householder responses<br />
to a survey<br />
Somewhat likely to<br />
somewhat unlikely;<br />
depends on student/<br />
householder responses<br />
to a survey<br />
Somewhat likely to<br />
somewhat unlikely;<br />
depends on student/<br />
householder responses<br />
to a survey<br />
Note. The Valid Indicator, Participation Rates, Minimal Nonresponse, <strong>an</strong>d E<strong>as</strong>y Access columns represent desirable characteristics of <strong>SES</strong> me<strong>as</strong>ures;<br />
each <strong>SES</strong> me<strong>as</strong>ure is evaluated on each characteristic. Results <strong>for</strong> householder education are approximately the same <strong>as</strong> those <strong>for</strong> householder occupation<br />
<strong>an</strong>d are not reported. Examples of the use of these me<strong>as</strong>ures are free or reduced price lunch eligibility (Stein et al., 2008); householder income,<br />
education, <strong>an</strong>d occupation (Hauser & Warren, 1997; Seyfried, 1998); number of siblings (H<strong>an</strong>ushek, 1992); <strong>an</strong>d home resources (Chen, Lee, &<br />
Stevenson, 1996; Martin, Mullis, & Chrostowski, 2004). Values in parentheses represent a r<strong>an</strong>ge of nonresponse rates b<strong>as</strong>ed on the Education<br />
Longitudinal Survey (National Center <strong>for</strong> Education Statistics, 2004); National Educational Longitudinal Study of 1988 (Hauser & Andrews, 2007);<br />
National Household Educational Surveys Program of 2007 (National Center <strong>for</strong> Education Statistics, 2009); National Longitudinal Survey of Youth (Cole<br />
& Currie, 1994); Trends in International Mathematics <strong>an</strong>d Science Study (Martin et al., 2004); Chen et al. (1996); Wardle, Robb, <strong>an</strong>d Johnson (2002);<br />
V<strong>an</strong> Horn, Bellis, <strong>an</strong>d Snyder (2001); <strong>an</strong>d Coscia et al. (2001).<br />
variables are not available then a home resources checklist could<br />
be a re<strong>as</strong>onable alternative.<br />
Second, in the presence of more signific<strong>an</strong>t nonresponse, the<br />
me<strong>as</strong>ures in Table 3, with the exception of the FRL <strong>an</strong>d number<br />
of siblings variables, are likely to show nonresponse bi<strong>as</strong>. For<br />
example, results from several large national studies (see Table 3)<br />
indicate that nonresponse rates <strong>for</strong> householder income r<strong>an</strong>ged<br />
between 12% <strong>an</strong>d 25%. Studies using a home resources me<strong>as</strong>ure<br />
showed even greater variation in nonresponse rates, r<strong>an</strong>ging<br />
between 1% <strong>an</strong>d 30%. A careful examination of the nonresponse<br />
patterns is likely to provide evidence of nonresponse bi<strong>as</strong>.<br />
Missing <strong>SES</strong> data are part of a broader missing data problem that<br />
h<strong>as</strong> received a good deal of attention in various literatures, including<br />
education (e.g., Allison, 2001; Little & Rubin, 2002; Peng, Harwell,<br />
Liou, & Ehm<strong>an</strong>, 2006). The literature on missing data distinguishes<br />
between three missing data c<strong>as</strong>es that are import<strong>an</strong>t in examining<br />
the implications of nonresponse <strong>for</strong> <strong>SES</strong> me<strong>as</strong>ures.<br />
If missing <strong>SES</strong> data are missing completely at r<strong>an</strong>dom<br />
(MCAR), the probability that <strong>an</strong> observation is missing does not<br />
maRch 2010 127
depend on the observed response or on the missing value. In this<br />
inst<strong>an</strong>ce, inferences b<strong>as</strong>ed on available data will not be bi<strong>as</strong>ed,<br />
although parameter estimation <strong>an</strong>d hypothesis testing may be<br />
affected by the loss of data (Little & Rubin, 2002). If missing<br />
data are missing at r<strong>an</strong>dom (MAR), the probability that <strong>an</strong> observation<br />
is missing depends only on one or more observed covariates.<br />
Under MAR, <strong>an</strong>alyzing available data produces unbi<strong>as</strong>ed<br />
<strong>an</strong>d efficient estimates if the observed covariates are included in<br />
the <strong>an</strong>alyses. Missing data that are not missing at r<strong>an</strong>dom<br />
(NMAR) me<strong>an</strong>s that the probability that <strong>an</strong> observation is missing<br />
depends on the missing data.<br />
Determining whether missing <strong>SES</strong> data are MCAR, MAR, or<br />
NMAR is frequently difficult <strong>an</strong>d relies heavily on a clear underst<strong>an</strong>ding<br />
of the facets of a study (e.g., sampling, treatment delivery,<br />
instrumentation, <strong>an</strong>d data collection). For missing <strong>SES</strong> data,<br />
MCAR seems unlikely in m<strong>an</strong>y c<strong>as</strong>es, where<strong>as</strong> MAR may be<br />
plausible with me<strong>as</strong>ures such <strong>as</strong> householder education <strong>an</strong>d occupation.<br />
In c<strong>as</strong>es where MCAR or MAR hold, multiple imputation<br />
c<strong>an</strong> be used to estimate missing <strong>SES</strong> values (e.g., householder<br />
education) using available data, eliminating nonresponse (see<br />
Allison, 2001; Little & Rubin, 2002; Peng et al., 2006). If missing<br />
data are NMAR, <strong>as</strong> might be expected with householder<br />
income because higher income households are less likely to report<br />
this in<strong>for</strong>mation, imputation methods are not recommended, <strong>as</strong><br />
they lead to bi<strong>as</strong>ed estimates <strong>an</strong>d inferences.<br />
Third, if signific<strong>an</strong>t nonresponse bi<strong>as</strong> is expected (perhaps due<br />
partly or largely to accessibility <strong>an</strong>d cost issues) the FRL variable<br />
may be the only alternative. However, a researcher opting to use<br />
FRL <strong>as</strong> <strong>an</strong> <strong>SES</strong> me<strong>as</strong>ure needs to recognize that FRL is a poor<br />
me<strong>as</strong>ure of a student’s access to economic resources <strong>an</strong>d to clearly<br />
identify the expected impact of deficiencies of this variable on<br />
study inferences (e.g., what is the impact on inferences if 20% of<br />
the sampled students are miscl<strong>as</strong>sified on the free lunch variable?).<br />
Conclusion<br />
The NSLP is deeply ingrained in K–12 education in the United<br />
States. An examination of the origins of the program shows that<br />
the intent w<strong>as</strong> to improve student nutrition, especially among the<br />
poorest students, which, it w<strong>as</strong> re<strong>as</strong>oned, would incre<strong>as</strong>e student<br />
learning. Available evidence generally supports the positive albeit<br />
small impact of this program on student nutrition <strong>an</strong>d learning.<br />
An import<strong>an</strong>t outcome of the growth of this program h<strong>as</strong><br />
been the use of eligibility <strong>for</strong> a free lunch <strong>as</strong> a me<strong>as</strong>ure of a student’s<br />
<strong>SES</strong> in education research. Perhaps the popularity of the<br />
free lunch variable <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong> c<strong>an</strong> be attributed to the<br />
use of federal poverty guidelines in determining student eligibility<br />
or the description of the free lunch variable <strong>as</strong> <strong>an</strong> indicator of<br />
economic adv<strong>an</strong>tage in the No Child Left Behind (2002) legislation.<br />
Or perhaps this popularity is simply due to the relatively<br />
e<strong>as</strong>y access to students’ free lunch status available to education<br />
researchers.<br />
Whatever the expl<strong>an</strong>ation, student eligibility <strong>for</strong> a free lunch<br />
continues to be used <strong>as</strong> a me<strong>as</strong>ure of <strong>SES</strong> in education research,<br />
including federally funded studies <strong>an</strong>d reports with import<strong>an</strong>t<br />
policy implications. This variable is typically used in data <strong>an</strong>alyses<br />
to statistically control <strong>for</strong> the effect of <strong>SES</strong> on educational<br />
outcomes, to incre<strong>as</strong>e statistical power, <strong>an</strong>d to enh<strong>an</strong>ce causality<br />
arguments. Despite its frequent use, however, the free lunch<br />
128<br />
educational ReseaRcheR<br />
variable possesses import<strong>an</strong>t deficiencies that suggest that other<br />
me<strong>as</strong>ures of <strong>SES</strong> should be considered.<br />
The process of identifying other me<strong>as</strong>ures ideally relies on a<br />
theoretical model of <strong>SES</strong> to conceptualize this construct <strong>an</strong>d<br />
select appropriate me<strong>as</strong>ures. Un<strong>for</strong>tunately, education researchers<br />
who turn to the <strong>SES</strong> literature <strong>for</strong> guid<strong>an</strong>ce will find signific<strong>an</strong>t<br />
disagreement on key definitions, terms, theoretical <strong>as</strong>sumptions,<br />
<strong>an</strong>d me<strong>as</strong>urement strategies. What is needed is work that connects<br />
theoretical models of <strong>SES</strong> <strong>an</strong>d strategies <strong>for</strong> me<strong>as</strong>uring <strong>SES</strong><br />
in educational contexts.<br />
An import<strong>an</strong>t practice <strong>for</strong> the education researcher to adopt is to<br />
carefully describe what <strong>SES</strong> is intended to represent in his or her<br />
study, <strong>for</strong> example, a student’s access to economic resources, <strong>an</strong>d to<br />
provide a clear rationale <strong>for</strong> selecting a me<strong>as</strong>ure of <strong>SES</strong> that is consistent<br />
with the study’s purpose. In c<strong>as</strong>es in which nonresponse is<br />
expected to be minimal, householder income, education, or occupation<br />
have much to recommend them <strong>as</strong> me<strong>as</strong>ures of a student’s <strong>SES</strong>.<br />
For signific<strong>an</strong>t nonresponse, education researchers may be<br />
able to turn to the literature on missing data <strong>for</strong> guid<strong>an</strong>ce. If missing<br />
values are <strong>as</strong>sumed to be MCAR or MAR, data imputation<br />
may be a plausible solution. In c<strong>as</strong>es where FRL status is the only<br />
alternative, it is import<strong>an</strong>t <strong>for</strong> researchers to recognize that this<br />
variable is a poor me<strong>as</strong>ure of a student’s access to economic<br />
resources. More generally, researchers need to clearly identify the<br />
expected impact of import<strong>an</strong>t deficiencies of <strong>an</strong>y variable selected<br />
to serve <strong>as</strong> a me<strong>as</strong>ure of student <strong>SES</strong> on study inferences.<br />
REFEREnCES<br />
Adler, N. E., Boyce, T., Chesney, M. A., Cohen, S., Folkm<strong>an</strong>, S., Kahn,<br />
R. L., et al. (1994). Socioeconomic status <strong>an</strong>d health: The challenge<br />
of the gradient. Americ<strong>an</strong> Psychologist, 49, 15–24.<br />
Alex<strong>an</strong>der, K. L., Entwisle, D. R., & Olson, L. S. (2007). L<strong>as</strong>ting consequences<br />
of the summer learning gap. Americ<strong>an</strong> Sociological Review,<br />
72, 167–180.<br />
Allison, P. D. (2001). Missing data. Thous<strong>an</strong>d Oaks, CA: Sage.<br />
Americ<strong>an</strong> Psychological Association. (2007a). Office on Socioeconomic<br />
Status. Retrieved September 18, 2008, from http://www.apa.org/pi/<br />
ses/homepage.html<br />
Americ<strong>an</strong> Psychological Association. (2007b). Report of the APA T<strong>as</strong>k<br />
Force on Socioeconomic Status. W<strong>as</strong>hington, DC: Author. Retrieved<br />
September 18, 2008, from http://www2.apa.org/pi/<strong>SES</strong>_t<strong>as</strong>k_<strong>for</strong>ce<br />
_report.pdf<br />
Annual update of the HHS poverty guidelines. (2008). Federal Register,<br />
73(15), 3971–3972.<br />
Blau, P. M., & Dunc<strong>an</strong>, O. D. (1967). The Americ<strong>an</strong> occupation structure.<br />
New York: John Wiley.<br />
Bollen, K. A., Gl<strong>an</strong>ville, J. L., & Stecklov, G. (2001). Socioeconomic<br />
status <strong>an</strong>d cl<strong>as</strong>s in studies of fertility <strong>an</strong>d health in developing countries.<br />
Annual Review of Sociology, 27, 153–185.<br />
Borm<strong>an</strong>, G. D., & Hewes, G. M. (2002). The long-term effects <strong>an</strong>d<br />
cost-effectiveness of success <strong>for</strong> all. Educational Evaluation <strong>an</strong>d Policy<br />
Analysis, 24, 243–266.<br />
Boykin, A. W., Colem<strong>an</strong>, S. T., Lilja, A. J., & Tyler, K. M. (2004).<br />
Research vist<strong>as</strong> in the communal learning paradigm (Center <strong>for</strong> Research<br />
on the Education of <strong>Student</strong>s Placed at Risk Rep. No. 68). Retrieved<br />
November 10, 2008, from http://www.csos.jhu.edu/crespar/tech<br />
Reports/ Report68.pdf<br />
Breen, R. (2005). Foundations of cl<strong>as</strong>s <strong>an</strong>alysis in the Weberi<strong>an</strong> tradition.<br />
In E. O. Wright (Ed.), Alternative foundations of cl<strong>as</strong>s <strong>an</strong>alysis<br />
(pp. 41–69). Cambridge, UK: Cambridge University Press.
Bry<strong>an</strong>t, E. C., Glazer, E., H<strong>an</strong>sen, M. A., & Kirsch, A. (1974).<br />
Associations between educational outcomes <strong>an</strong>d background variables<br />
[Monograph]. Denver, CO: National Assessment of Educational<br />
Progress.<br />
Burkhead, J., Fox, T. G., & Holl<strong>an</strong>d, J. W. (1967). Input <strong>an</strong>d output in<br />
large city high schools. Syracuse, NY: Syracuse University Press.<br />
Campbell, R. T. (1983). Subst<strong>an</strong>tive <strong>an</strong>d statistical considerations in the<br />
interpretation of multiple me<strong>as</strong>ures of <strong>SES</strong>. Social Forces, 62, 450–<br />
466.<br />
Chen, C., Lee, S. Y., & Stevenson, H. W. (1996). Long-term prediction<br />
of academic achievement of Americ<strong>an</strong>, Chinese, <strong>an</strong>d Jap<strong>an</strong>ese adolescents.<br />
Journal of Educational Psychology, 88, 750–759.<br />
Cirino, P. T., Chin, C. E., Sevcik, R. A., Wolf, M., Lovett, M., & Morris,<br />
R. D. (2002). Me<strong>as</strong>uring socioeconomic status: Reliability <strong>an</strong>d preliminary<br />
validity <strong>for</strong> different approaches. Assessment, 9, 145–155.<br />
Cole, N., & Currie, J. (1994). Reported income in the NLSY: Consistency<br />
checks <strong>an</strong>d methods <strong>for</strong> cle<strong>an</strong>ing the data (Working Paper No. T0160).<br />
Cambridge, MA: National Bureau of Economic Research.<br />
Colem<strong>an</strong>, J. S., Campbell, E. Q., Hobson, C. J., McPartl<strong>an</strong>d, J.,<br />
Mood, A. M., Weinfield, F. D., et al. (1966). Equality of educational<br />
opportunity (2 vols.). W<strong>as</strong>hington, DC: U.S. Government Printing<br />
Office.<br />
Coscia, J. M., Christensen, B. K., Henry, R. R., Wallston, K., Radcliffe,<br />
J., & Rutstein, R. (2001). Effects of home environment, socioeconomic<br />
status, <strong>an</strong>d health status on cognitive functioning in children<br />
with HIV-1 infection. Journal of Pediatric Psychology, 26, 321–329.<br />
Dauber, S. L., Alex<strong>an</strong>der, K. L., & Entwisle, D. R. (1996). Tracking <strong>an</strong>d<br />
tr<strong>an</strong>sitions through the middle grades: Ch<strong>an</strong>neling educational trajectories.<br />
Sociology of Education, 69, 290–307.<br />
Dev<strong>an</strong>ey, B. L., Ellwood, M. R., & Love, J. M. (1997). Programs that<br />
mitigate the effects of poverty on children. Future of Children, 7,<br />
88–112.<br />
Economic Research Service. (2003). Direct certification in the National<br />
School <strong>Lunch</strong> Program: Impacts on program access <strong>an</strong>d integrity (Rep. No.<br />
03–009, prepared by P. Gle<strong>as</strong>on, T. T<strong>as</strong>se, K. Jackson, & P. Memeth).<br />
W<strong>as</strong>hington, DC: U.S. Department of Agriculture. Retrieved<br />
October 11, 2008, from http://www.ers.usda.gov/publications/<br />
ef<strong>an</strong>03009/ef<strong>an</strong>03009.pdf<br />
Economic Research Service. (2006). Profiles of particip<strong>an</strong>ts in the<br />
National School <strong>Lunch</strong> Program: Data from two national surveys (Rep.<br />
No. EIB 17, prepared by C. Newm<strong>an</strong> & K. Ralston). Alex<strong>an</strong>dria, VA:<br />
U.S. Department of Agriculture. Retrieved April 5, 2008, from www<br />
.ers.usda.gov/publications/eib17/eib17.pdf<br />
Economic Research Service. (2008). The National School <strong>Lunch</strong> Program:<br />
Background, trends, <strong>an</strong>d issues (Rep. No. 61, prepared by K. Ralston,<br />
C. Newm<strong>an</strong>, A. Clauson, J. Guthrie, & J. Buzby). Alex<strong>an</strong>dria, VA:<br />
U.S. Department of Agriculture. Retrieved September 26, 2008,<br />
from www.ers.usda.gov/Publications/ERR61/ERR61.pdf<br />
Edwards-Hewitt, T., & Gray, J. J. (1995). Comparison of me<strong>as</strong>ures of<br />
socioeconomic status between ethnic groups. Psychological Reports, 77,<br />
699–702.<br />
Entwisle, D. R., & Astone, N. M. (1994). Some practical guidelines <strong>for</strong><br />
me<strong>as</strong>uring youth’s race/ethnicity <strong>an</strong>d socioeconomic status. Child<br />
Development, 65, 1521–1540.<br />
Food <strong>an</strong>d Nutrition Service. (1990). Study of Income Verification in the<br />
National School <strong>Lunch</strong> Program, Final report (Vol. 1, Rep. No. FNS-<br />
53–3198-6–44, prepared by R. St. Pierre, M. Puma, M. Battaglia, &<br />
J. Layzer). Alex<strong>an</strong>dria, VA: U.S. Department of Agriculture. Retrieved<br />
October 11, 2008, from http://www.fns.usda.gov/o<strong>an</strong>e/MENU/<br />
Published/CNP/cnp-archive.htm<br />
Food <strong>an</strong>d Nutrition Service. (2008). Program fact sheet. Alex<strong>an</strong>dria, VA:<br />
U.S. Department of Agriculture. Retrieved September 20, 2008,<br />
from http://www.fns.usda.gov/cnd/<strong>Lunch</strong>/<br />
General Accounting Office. (2003). Public Schools: Comparison of<br />
achievement results <strong>for</strong> students attending privately m<strong>an</strong>aged <strong>an</strong>d<br />
traditional schools in six cities (Rep. No. GAO-04–62, prepared by D.<br />
Edwards & P. Elston). Retrieved October 11, 2008, from http://www.<br />
gao.gov/new.items/d0462.pdf<br />
Gle<strong>as</strong>on, P. (2008). Direct certification in the National School <strong>Lunch</strong><br />
Program exp<strong>an</strong>ds access <strong>for</strong> children. Journal of Policy Analysis <strong>an</strong>d<br />
M<strong>an</strong>agement, 27, 82–103.<br />
Groves, R. M. (2006). Nonresponse rates <strong>an</strong>d nonresponse bi<strong>as</strong> in<br />
household surveys. Public Opinion Quarterly, 70, 646–675.<br />
Grusky, G. (2005). Foundations of Durkheimi<strong>an</strong> cl<strong>as</strong>s <strong>an</strong>alysis. In E. O.<br />
Wright (Ed.), Alternative foundations of cl<strong>as</strong>s <strong>an</strong>alysis (pp. 70–118).<br />
Cambridge, UK: Cambridge University Press.<br />
Gunderson, G. W. (2003). The National School <strong>Lunch</strong> Program:<br />
Background <strong>an</strong>d development. New York: Nova Science.<br />
H<strong>an</strong>ushek, E. (1992). The trade-off between child qu<strong>an</strong>tity <strong>an</strong>d quality.<br />
Journal of Political Economy, 100, 84–117.<br />
Harrington, A., Marshall, B. L., & Muller, H. P. (Eds.). (2006).<br />
Encyclopedia of social theory. New York: Routledge.<br />
Harwell, M. R., Maeda, Y., & Lee, K. (2004, April). Replicating <strong>an</strong>d<br />
extending White’s (1982) Meta-<strong>an</strong>alysis of the relationship between <strong>SES</strong><br />
<strong>an</strong>d student achievement. Paper presented at the <strong>an</strong>nual meeting of the<br />
Americ<strong>an</strong> Educational Research Association, S<strong>an</strong> Diego, CA.<br />
Haug, M. R. (1977). Me<strong>as</strong>urement in social stratification. Annual<br />
Review of Sociology, 3, 51–77.<br />
Hauser, R. M. (1994). Me<strong>as</strong>uring socioeconomic status in studies of<br />
child development. Child Development, 65, 1541–1545.<br />
Hauser, R. M., & Andrew, M. (2007). Reliability of student <strong>an</strong>d parent<br />
reports of socioeconomic status in NELS-88. Working paper presented<br />
at ITP seminar at the University of Wisconsin–Madison. Retrieved<br />
August 10, 2009, from http://www.wcer.wisc.edu/itp/Spring%20<br />
08%20seminar/HauserNELS-<strong>SES</strong>%20me<strong>as</strong>urement_070607a.pdf<br />
Hauser, R. M., & Warren, J. R. (1997). Sociological indexes <strong>for</strong> occupations:<br />
A review, update, <strong>an</strong>d critique. Sociological Methodology, 27, 177–298.<br />
Herr, A., & Tobi<strong>as</strong>, S. (1970). Achievement via programmed instruction<br />
<strong>an</strong>d socioeconomic status. Psychology in the Schools, 1, 53–56.<br />
Hobbs, G., & Vignoles, A. (2007). Is free school meal status a valid proxy<br />
<strong>for</strong> socio-economic status (in schools research)? (London School of<br />
Economics, Centre <strong>for</strong> the Economics of Education, Discussion<br />
Paper No. 0084). Retrieved November 29, 2008, from http://cee.lse<br />
.ac.uk/cee%20dps/ceedp84.pdf<br />
Holley, C. E. (1916). The relationship between persistence in school <strong>an</strong>d<br />
home conditions. Chicago: University of Chicago Press.<br />
Internal Revenue Service. (2008). Earned income tax credit (EITC) (Pub.<br />
No. 596). Retrieved December 8, 2008, from http://www.irs.gov/<br />
pub/irs-pdf/p596.pdf<br />
Jones, F. L., & McMill<strong>an</strong>, J. (2001). Scoring occupational categories <strong>for</strong><br />
social research: A review of current practice, with Australi<strong>an</strong> examples.<br />
Work, Employment, <strong>an</strong>d Society, 51, 539–563.<br />
Kerckhoff, A. C. (1975). Antecedents of academic per<strong>for</strong>m<strong>an</strong>ce <strong>an</strong>d educational<br />
attainment (Final Rep., Project B-226). Chicago: Spencer<br />
Foundation.<br />
Kurki, A., Boyle, A., & Aladjem, D. K. (2005, April). Beyond free lunch:<br />
Alternative poverty me<strong>as</strong>ures in educational research <strong>an</strong>d program evaluation.<br />
Paper presented at the <strong>an</strong>nual meeting of the Americ<strong>an</strong><br />
Educational Research Association, Montreal, C<strong>an</strong>ada.<br />
Levine, S. (2008). School lunch politics: The surprising history of<br />
America’s favorite welfare program. Princeton, NJ: Princeton University<br />
Press.<br />
Liew, J., McTigue, E. M., Barrios, L., & Hughes, J. N. (2008). Adaptive<br />
<strong>an</strong>d ef<strong>for</strong>tful control <strong>an</strong>d academic self-efficacy beliefs on achievement:<br />
A longitudinal study of 1st through 3rd graders. Early Childhood<br />
Research Quarterly, 23, 515–526.<br />
maRch 2010 129
Little, R. J. A., & Rubin, D. B. (2002). Statistical <strong>an</strong>alysis with missing<br />
data (2nd ed.) New York: John Wiley.<br />
Lubienski, S. T., & Lubienski, C. (2006). School sector <strong>an</strong>d academic<br />
achievement: A multilevel <strong>an</strong>alysis of NAEP mathematics data.<br />
Americ<strong>an</strong> Educational Research Journal, 43, 651–698.<br />
Lynd, R. S., & Lynd, H. M. (1929). Middletown: A study in Americ<strong>an</strong><br />
culture. New York: Harcourt Brace.<br />
Mathematica Policy Research. (2005). Trends in nutrition policy (Issue<br />
Brief No. 2, Rep. No. PP05–14, prepared by P. Gle<strong>as</strong>on & J. Burghardt).<br />
Princeton, NJ: Author. Retrieved October 11, 2008, from http://www<br />
.mathematica-mpr.com/publications/PDFs/nslpensuring.pdf<br />
Mathematica Policy Research. (2008). High school re<strong>for</strong>m in Boston public<br />
schools: The effect of focus on high schools on student academic outcomes<br />
(Rep. No. 8927- 00, prepared by S. James-Burdumy, I. Perez-Johnson, &<br />
S. Vartivari<strong>an</strong>). Retrieved December 11, 2008, from https://www.policy<br />
archive.org/bitstream/h<strong>an</strong>dle/10207/8803/HighSchoolRe<strong>for</strong>m .pdf?<br />
sequence=1<br />
May, H. (2002). Development <strong>an</strong>d evaluation of <strong>an</strong> internationally comparable<br />
scale of student socioeconomic status using survey data from<br />
TIMSS. Unpublished doctoral dissertation, University of Pennsylv<strong>an</strong>ia<br />
(AAT 3073031).<br />
Me<strong>an</strong>s, G. H., Me<strong>an</strong>s, R. S., Osborne, J. L., & Elsome, B. F. (1973).<br />
T<strong>as</strong>k persistence <strong>as</strong> a function of verbal rein<strong>for</strong>cement <strong>an</strong>d socioeconomic<br />
status. Cali<strong>for</strong>nia Journal of Educational Research, 24,<br />
5–11.<br />
Miner, B. (1968). Sociological background variables affecting school<br />
achievement. Journal of Educational Research, 61, 372–381.<br />
Molnar, A., Smith, P., Zahorik, J., Palmer, Halbach, A., & Ehrle, K.<br />
(1999). Evaluating the SAGE program: A pilot program in targeted<br />
pupil–teacher reduction in Wisconsin. Educational Evaluation <strong>an</strong>d<br />
Policy Analysis, 21, 165–177.<br />
Montague, D. O. (1964). Arithmetic concepts of kindergarten children<br />
in contr<strong>as</strong>ting socioeconomic are<strong>as</strong>. Elementary School Journal, 64,<br />
393–397.<br />
Nam, C. B., & Powers, M. G. (1983). The socioeconomic approach to<br />
status me<strong>as</strong>urement (with a guide to occupational <strong>an</strong>d socioeconomic status<br />
scores). Houston, TX: Cap <strong>an</strong>d Gown Press.<br />
National Center <strong>for</strong> Education Statistics. (2003–2004). Public elementary/secondary<br />
school universe survey (Rep. No. NCES 2006–324, prepared<br />
by J. Sable, J. Thom<strong>as</strong>, & Q. Shen). W<strong>as</strong>hington, DC: U.S.<br />
Department of Education. Retrieved September 20, 2008, from http://<br />
eric.ed.gov/ERICDocs/data/ericdocs2sql/content_storage_01/0000<br />
019b/ 80/29/df/13.pdf<br />
National Center <strong>for</strong> Education Statistics. (2004). Education Longitudinal<br />
Study of 2002: B<strong>as</strong>e year data file user’s m<strong>an</strong>ual (Rep. No. 2004–405,<br />
prepared by S. Ingels, D. Pratt, J. Rogers, P. Siegel, & E. Stutts).<br />
W<strong>as</strong>hington, DC: U.S. Department of Education.<br />
National Center <strong>for</strong> Education Statistics. (2006a). A closer look at charter<br />
schools using hierarchical linear modeling. (Rep. No. 2006–460, prepared<br />
by H. Braun, F. Jenkins, & W. Grigg). W<strong>as</strong>hington, DC: U.S.<br />
Department of Education. Retrieved October 11, 2008, from http://<br />
nces.ed.gov/nationsreportcard/pdf/studies/2006460.pdf<br />
National Center <strong>for</strong> Education Statistics. (2006b). United States high<br />
school sophomores: A twenty-two year comparison, 1980–2000 (Rep.<br />
No. NCES 2006–327, prepared by M. Cahal<strong>an</strong>, S. Ingels, L. Burns,<br />
M. Pl<strong>an</strong>ty, & B. D<strong>an</strong>iel). W<strong>as</strong>hington, DC: U.S. Department of<br />
Education. Retrieved September 20, 2008, from http://nces.ed.gov/<br />
pubs2006/2006327.pdf<br />
National Center <strong>for</strong> Education Statistics. (2007). The nation’s report<br />
card: Mathematics (Rep. No. NCES 2007–494, prepared by J. Lee,<br />
W. Grigg, & G. Dion). W<strong>as</strong>hington, DC: U.S. Department of Education.<br />
Retrieved September 20, 2008, from http://nces.ed.gov/nationsre<br />
portcard/pdf/main2007/2007494.pdf<br />
130<br />
educational ReseaRcheR<br />
National Center <strong>for</strong> Education Statistics. (2008a). The condition of<br />
education 2008 (Rep. No. NCES 2008–031, prepared by M. Pl<strong>an</strong>ty,<br />
W. Hussar, T. Snyder, S. Prov<strong>as</strong>nik, G. Kena, R. Dinkes, et al.).<br />
W<strong>as</strong>hington, DC: U.S. Department of Education. Retrieved<br />
September 20, 2008 from http://nces.ed.gov/pubs2008/2008031.pdf<br />
National Center <strong>for</strong> Education Statistics. (2008b). Toward a new me<strong>as</strong>ure<br />
of socioeconomic status in NAEP. Retrieved June 10, 2009, from<br />
http://74.125.45.132/search?q=cache:Bsz_YfvfOdIJ:www.ccsso.org/<br />
content/pdfs/NAEP%2520<strong>SES</strong>%2520Research.ppt+number+of+sib<br />
lings+<strong>as</strong>+ses+me<strong>as</strong>ure&cd=1&hl=en&ct=clnk&gl=us&client =fire<br />
fox-a<br />
National Center <strong>for</strong> Education Statistics. (2008c). Trends among high<br />
school seniors, 1972–2004 (Rep. No. NCES 2008–320, prepared by<br />
S. Ingels & B. Dalton). W<strong>as</strong>hington, DC: U.S. Department of<br />
Education. Retrieved December 5, 2008, from http://nces.ed.gov/<br />
pubs2008/2008320.pdf<br />
National Center <strong>for</strong> Education Statistics. (2009). National Household<br />
Education Surveys Program of 2007: Methodology report (Rep. No.<br />
2009–029, prepared by M. Hagedorn, S. Roth, P. Carver, W. V<strong>an</strong> de<br />
Kerckhove, & S. Smith). W<strong>as</strong>hington, DC: U.S. Department of<br />
Education. Retrieved August 10, 2009, from http://nces.ed.gov/<br />
pubs2009/2009047_1.pdf<br />
National School <strong>Lunch</strong> Program. (2008). Fact sheet. Retrieved September<br />
16, 2008, from http://www.fns.usda.gov/cnd/govern<strong>an</strong>ce/notices/<br />
naps/NAPs.htm<br />
Nichols, S. L., & Berliner, D. C. (2008). Collateral damage: How highstakes<br />
testing corrupts America’s schools. Cambridge, MA: Harvard<br />
Education Press.<br />
Nierm<strong>an</strong>, G. E., & Veak, M. H. (1997). Effect on selected recruiting<br />
strategies on beginning instrumentalists’ participation decisions.<br />
Journal of Research in Music Education, 45, 380–389.<br />
No Child Left Behind Act of 2001. Pub. L. No. 107-110, 115 Stat. 1425<br />
(2002).<br />
Oakes, J. M., & Rossi, P. H. (2003). The me<strong>as</strong>urement of <strong>SES</strong> in health<br />
research: Current practice <strong>an</strong>d steps toward a new approach. Social<br />
Science <strong>an</strong>d Medicine, 56, 769–784.<br />
Office of Inspector General. (1997). Food <strong>an</strong>d Consumer Service,<br />
National School <strong>Lunch</strong> Program verification of applications in Illinois<br />
(Rep. No. 27010–0011-Ch). Alex<strong>an</strong>dria, VA: U.S. Department of<br />
Agriculture.<br />
Office of Research, Nutrition, <strong>an</strong>d Analysis. (1994). School lunch eligible<br />
non-particip<strong>an</strong>ts: Final report (Rep. No. 53–3198–018, prepared by F.<br />
Gl<strong>an</strong>tz, R. Berg, D. Porcari, E. Sackoff, & S. Pazer). Alex<strong>an</strong>dria, VA: U.S.<br />
Department of Agriculture. Retrieved October 11, 2008, from http://<br />
www.fns.usda.gov/o<strong>an</strong>e/MENU/Published/CNP/cnp-archive.htm<br />
Office of Research, Nutrition, <strong>an</strong>d Analysis. (2004). Evaluation of<br />
the National School <strong>Lunch</strong> Program Application/Verification Pilot<br />
Projects: Vol. 1. Impacts on deterrence, barriers, <strong>an</strong>d accuracy (Special<br />
Nutrition Program Report Series, Rep. No. CN-04-AV5, prepared by<br />
J. Burghardt, P. Gle<strong>as</strong>on, M. Sinclair, R. Cohen, L. Hulsey, &<br />
J. Milliner-Waddell). W<strong>as</strong>hington, DC. U.S. Department of<br />
Agriculture. Retrieved October 11, 2008, from http://www.fns.usda.<br />
gov/o<strong>an</strong>e/MENU/Published/CNP/FILES/NSLPPilotVol1.pdf<br />
Office of Research, Nutrition, <strong>an</strong>d Analysis. (2007). NSLP/SBP Access,<br />
Participation, <strong>Eligibility</strong>, <strong>an</strong>d Certification Study—Erroneous payments<br />
in the NSLP <strong>an</strong>d SBP: Vol. 1. Study findings (Rep. No. CN-07-APEC,<br />
prepared by M. Ponza et al.). Alex<strong>an</strong>dria, VA: U.S. Department of<br />
Agriculture. Retrieved October 11, 2008, from http://www.fns.usda.<br />
gov/OANE/MENU/Published/CNP/FILES/apecvol1.pdf<br />
Orsh<strong>an</strong>sky, M. (1965). Counting the poor: Another look at the poverty<br />
profile. Social Security Bulletin, 28, 3–29.<br />
Payne, K. J., & Biddle, B. J. (1999). Poor school funding, child poverty,<br />
<strong>an</strong>d mathematics achievement. Educational Researcher, 28(6), 4–13.
Peng, C. Y., Harwell, M. R., Liou, S. M., & Ehm<strong>an</strong>, L. H. (2006).<br />
Adv<strong>an</strong>ces in missing data methods <strong>an</strong>d implications <strong>for</strong> educational<br />
research. In S. S. Sawilowsky (Ed.), Real data <strong>an</strong>alysis (pp. 31–78).<br />
Charlotte, NC: In<strong>for</strong>mation Age.<br />
Pentz, M. A., Trebow, E. A., H<strong>an</strong>sen, W. B., MacKinnon, D. P., Dwyer,<br />
J. H., Johnson, C. A., et al. (1990). Effects of program implementation<br />
on adolescent drug use behavior: The Midwestern Prevention<br />
Project (MPP). Evaluation Review, 14, 264–289.<br />
Powers, M. G. (1982). Me<strong>as</strong>ures of socioeconomic status: An introduction.<br />
In M. G. Powers (Ed.), Me<strong>as</strong>ures of socioeconomic status: Current<br />
issues (pp. 1–28). Boulder, CO: Westview.<br />
Purves, A. C. (1973). Literature education in ten countries: International<br />
studies in education. Stockholm, Sweden: Almquist & Wiksell.<br />
Raymond, R. (1968). Determin<strong>an</strong>ts of the quality of primary <strong>an</strong>d secondary<br />
public education in West Virginia. Journal of Hum<strong>an</strong> Resources,<br />
3, 450–470.<br />
Richard B. Russell National School <strong>Lunch</strong> Act of 1946, 42 U.S.C. 1751<br />
et seq.<br />
Rose, D., & Pevalin, D. J. (2003). The NS-SEC explained. In D. Rose<br />
& D. J. Pevalin (Eds.), A researcher’s guide to the national statistics socioeconomic<br />
cl<strong>as</strong>sification (pp. 28–43). London: Sage.<br />
Seyfried, S. F. (1998). Academic achievement of Afric<strong>an</strong> Americ<strong>an</strong> preadolesc<strong>an</strong>ts:<br />
The influence of teacher perceptions. Americ<strong>an</strong> Journal of<br />
Community Psychology, 26, 381–402.<br />
Sirin, S. R. (2005). Socioeconomic status <strong>an</strong>d academic achievement: A<br />
meta-<strong>an</strong>alytic review of research. Review of Educational Research, 75,<br />
417–453.<br />
Spaeth, J. L. (1976). Cognitive complexity: A dimension underlying the<br />
socioeconomic achievement process. In W. H. Sewall, R. M. Hauser,<br />
& D. L. Featherm<strong>an</strong> (Eds.), Schooling <strong>an</strong>d achievement in Americ<strong>an</strong><br />
society (pp. 103–131). New York: Academic Press.<br />
Spellings, M. (2007). Building on results: A blueprint <strong>for</strong> strengthening the<br />
No Child Left Behind Act. W<strong>as</strong>hington, DC: U.S. Department of<br />
Education. Retrieved November 11, 2008, from http://www.ed.gov/<br />
policy/elsec/leg/nclb/buildingonresults.pdf<br />
St<strong>an</strong>fiel, J. D. (1973). Socioeconomic status <strong>as</strong> related to aptitude, attrition,<br />
<strong>an</strong>d achievement of college students. Sociology of Education, 46, 480–488.<br />
Stein, M. L., Berends, M., Fuchs, D., McM<strong>as</strong>ter, K., Saenz, L., Yen, L.,<br />
et al. (2008). Scaling up <strong>an</strong> early reading program: Relationships<br />
among teacher support, fidelity of implementation, <strong>an</strong>d student per<strong>for</strong>m<strong>an</strong>ce<br />
across different sites <strong>an</strong>d years. Educational Evaluation <strong>an</strong>d<br />
Policy Analysis, 30, 368–388.<br />
Martin, M. O., Mullis, I. V. S., & Chrostowski, S. J. (Eds.). (2004).<br />
TIMSS 2003 Technical Report: Findings from IEA’s Trends in<br />
International Mathematics <strong>an</strong>d Science Study at the fourth <strong>an</strong>d eighth<br />
grades. Chestnut Hill, MA: TIMSS & PIRLS International Study<br />
Center, Lynch School of Education, Boston College. Retrieved<br />
November 11, 2008, from http://timss.bc.edu/PDF/t03_download/<br />
T03TECHRPT.pdf<br />
V<strong>an</strong> Horn, M. L., Bellis, J. M., & Snyder, S. W. (2001). Family Resource<br />
Scale–Revised: Psychometrics <strong>an</strong>d validation of a me<strong>as</strong>ure of family<br />
resources in a sample of low-income families. Journal of Psychoeducational<br />
Assessment, 19, 54–68.<br />
Walpole, M. (2003). Socioeconomic status <strong>an</strong>d college: How <strong>SES</strong> affects<br />
college experiences <strong>an</strong>d outcomes. Review of Higher Education, 27,<br />
45–73.<br />
W<strong>an</strong>g, J., Niemi, D., & W<strong>an</strong>g, H. (2007). Impact of different per<strong>for</strong>m<strong>an</strong>ce<br />
<strong>as</strong>sessment cut scores on student promotion (Center <strong>for</strong> Research on<br />
Evaluation, St<strong>an</strong>dards, <strong>an</strong>d <strong>Student</strong> Testing, UCLA, Rep. No. 719).<br />
Retrieved November 10, 2008, from http://www.cse.ucla.edu/products/<br />
reports/R719.pdf<br />
Wardle, J., Robb, K., & Johnson, F. (2002). Assessing socioeconomic<br />
status in adolescents: The validity of a home affluence scale. Journal of<br />
Epidemiology <strong>an</strong>d Community Health, 56, 595–599.<br />
White, K. R. (1982). The relation between socioeconomic status <strong>an</strong>d<br />
academic achievement. Psychological Bulletin, 91, 461–481.<br />
Worley, S. E., & Story, W. E. (1967). Socioeconomic status <strong>an</strong>d l<strong>an</strong>guage<br />
facility of beginning first graders. Reading Teacher, 20, 400–403.<br />
Wright, E. O. (2002). Foundations of cl<strong>as</strong>s <strong>an</strong>alysis in the Marxist tradition.<br />
In E. O. Wright (Ed.), Alternative foundations of cl<strong>as</strong>s <strong>an</strong>alysis<br />
(pp. 6–40). Cambridge, UK: Cambridge University Press.<br />
AUTHORS<br />
MICHAEL HARWELL is a professor of educational psychology at the<br />
University of Minnesota, Department of Educational Psychology, 56 E<strong>as</strong>t<br />
River Road, Minneapolis, MN 55455; harwe001@umn.edu. His research<br />
interests include defining <strong>an</strong>d me<strong>as</strong>uring poverty in education, meta<strong>an</strong>alysis,<br />
<strong>an</strong>d the relationship between high school mathematics curricula<br />
<strong>an</strong>d college mathematics per<strong>for</strong>m<strong>an</strong>ce.<br />
BRANDON LeBEAU is a doctoral student at the University of<br />
Minnesota, Department of Educational Psychology, 56 E<strong>as</strong>t River Road,<br />
Minneapolis, MN 55455; lebea027@umn.edu. His research focuses on<br />
longitudinal <strong>an</strong>alysis <strong>an</strong>d characteristics of the National School <strong>Lunch</strong><br />
Program in educational contexts.<br />
M<strong>an</strong>uscript received February 25, 2009<br />
Revisions received July 15, 2009, <strong>an</strong>d August 19, 2009<br />
Accepted August 20, 2009<br />
maRch 2010 131