IN THE PAST DECADE, a dramatic increase in investments

(1)

RESEARCH BRIEF #9

Assessment Tools for Language and Literacy Development of

Young Dual Language Learners (DLLs)

Center for

Early Care and Education Research

Dual Language Learners

I

N THE PAST DECADE, a dramatic increase in investments in early childhood programs (Barnett et al., 2011) has been accompanied by an increase in the number of dual language learners (DLLs)—chil-

GUHQIURPELUWKWRDJH¿YH\HDUVZKRDUHJURZLQJXS LQKRPHVZKHUHODQJXDJHVRWKHUWKDQRULQDGGLWLRQWR

English are spoken (Aud et al., 2012; Aikens et al., 2011; Vogel et al., 2011). For example, recent reports from the Family and Child Experiences Survey (FACES) 2009 cohort, the ongoing national study of children in Head Start, indicate that more than 31 percent of preschool-

HUVLQWKHSURJUDPOLYHLQKRPHVZKHUHDODQJXDJHRWKHU

than English is spoken (Aikens et al., 2011). This is also true of almost one-third of children in Early Head Start nationally (Vogel et al., 2011). Reliable and valid assessments for this population are needed for evaluations of

HDUO\FKLOGKRRGSURJUDPVDQGKRZZHOOWKH\DUHPHHW- ing the needs of DLLs.

All children need assessments that are fair, equitable,

DQGZHOOFRQVWUXFWHG7KHUHVXOWVVKRXOGUHSUHVHQW VNLOOVNQRZOHGJHDQGEHKDYLRUVRI'//VWKDWDUHLP- portant for their development and later school success. Since the questions and probes for most assessments require the use of a single language, either by the assessor or the child, the results from an assessment of a child

ZKRLVOHDUQLQJWZRODQJXDJHVPD\QRWUHSUHVHQWIXOO\ KLVRUKHUNQRZOHGJHDQGVNLOOV

Examination of reliability and validity evidence helps us

VHOHFWPHDVXUHVE\SURYLGLQJLQIRUPDWLRQDERXWZKHQ DQGZLWKZKRPWKHDVVHVVPHQWZLOOSURYLGHWUXVWZRUWK\

and meaningful results. Information about reliability

FDQWHOOXVKRZGHSHQGDEOHWKHUHVXOWVDUHZKLOHHYL- GHQFHRIYDOLGLW\LQGLFDWHVZKHWKHUWKHDVVHVVPHQWLV PHDVXULQJZKDWLWLVVXSSRVHGWREHPHDVXULQJDQG ZKDWW\SHVRILQIHUHQFHVFDQEHGUDZQIURPWKHUHVXOWV %RWKUHOLDELOLW\DQGYDOLGLW\LQIRUPZKDWW\SHVRILQIHU- HQFHVFDQEHGUDZQIURPWKHUHVXOWV

Typically, information about evidence of the reliability and validity of particular assessments is found in the technical manuals that accompany them (see Table 1 for examples).

2WKHUWKDQDIHZGHVLJQHGVSHFL¿FDOO\IRU'//VPRVWDV- sessments report about the reliability and validity evidence based on samples that are more representative of children

ZKRDUHPRQROLQJXDODQGPDNHXSWKHPDMRULW\RIWKH

standardization samples. Test development is expensive, and publishers usually do not invest in seeking evidence of validity for subgroups such as DLLs.

Assessments are validated by accumulating evidence in relation to different types of inferences. An assessment may be valid for a particular purpose or representative of the skills of one group of children but not another.

7KHNH\LVVXHLVZKHWKHUDQDVVHVVPHQWUHDOO\PHDVXUHV ZKDWLWGHFODUHVWRPHDVXUHZKHQXVHGZLWKDJURXS

of children similar to those in the study and for that particular purpose (for example, program evaluation,

UHVHDUFKDERXWFKLOGUHQ¶VGHYHORSPHQWRULGHQWL¿FD- tion of a child’s need for more intensive instruction or specialized intervention).

Different types of evidence can be collected to support interpretation of an assessment and the results ob- tained from it. The normative group (the sample used to develop comparison scores on the assessment) is a key component. An assessment may be valid for comparing

FKLOGUHQIURPWKHVDPHOLQJXLVWLFJURXSEXWXQIDLUZKHQ

(2)

A measure may include items that function in differen

ZD\VDFURVVOLQJXLVWLFJURXSVLQFOXGLQJGLIIHUHQFHVLQ GLI¿FXOW\GLVFULPLQDWLRQDQGRUIDFWRUORDGLQJV7KH

abilities of one group may not be validly represented by the scores derived from the items, redereing cross-group

FRPSDULVRQVXQIDLU,QDVLPLODUZD\DQDVVHVVPHQW

may not be valid for making critical decisions about children or the programs serving those children. For example, an English vocabulary or language assessment

ZRXOGEHDYDOLGLQGLFDWRURIZKHWKHUDFKLOGLVOHDUQLQJ

English, but using standard scores based on children

ZKRVSHDNRQO\(QJOLVKWRGHWHUPLQHLID'//KDVD ODQJXDJHGLVDELOLW\ZRXOGQRW

To extend the available information about assessments

XVHGZLWK'//VZHH[DPLQHGUHVHDUFKVWXGLHVSXE- lished in the last 10 years that included young DLLs.

7KLVEULHIGUDZVIURPDPRUHGHWDLOHGUHSRUWH[DPLQLQJ

language and literacy measures used in seven large- scale government studies and thirty research studies conducted in the past decade (Bandel et al., 2012).1_This brief highlights key issues noted in the report and em-

SKDVL]HVTXHVWLRQVWRFRQVLGHUZKHQHYDOXDWLQJWKHDS-

SURSULDWHQHVVRIDVVHVVPHQWWRROVXVHGZLWK'//VDQG

interpreting research that involves such assessments.

How reliable is the assessment?

Reliability information indicates to consumers of re-

VHDUFKWKHWUXVWZRUWKLQHVVRIWKH¿QGLQJVDQGUHVXOWV

can differ by sample and study. Poor reliability can limit the ability to detect change over time or associations

EHWZHHQFRQVWUXFWVIRUH[DPSOHEHWZHHQYRFDEXODU\ DQGHDUO\OLWHUDF\²SDUWLFXODUO\ZKHQWKHVDPSOHVL]HLV VPDOO²DQGFDQOHDGWRÀDZHGFRQFOXVLRQV0DQ\VWXGLHV LQRXUUHYLHZGLGQRWSURYLGHDQ\VWXG\VSHFL¿FHYL- dence of reliability but reported only the published evi-

GHQFHIURPWKHDVVHVVPHQWPDQXDO2XUUHYLHZVXJJHVWV

that more information is needed about the performance

RIPHDVXUHVZLWK'//V

Internal Consistency

The most commonly reported indicator of reliability is

LQWHUQDOFRQVLVWHQF\²WKDWLVKRZFRQVLVWHQWO\WKHLWHPV ZLWKLQDQDVVHVVPHQWPHD- sure the construct of inter-

HVW+RZHYHUPRVWRIWKH SHHUUHYLHZHGVWXGLHVGLG QRWLQFOXGHVWXG\VSHFL¿F

estimates (reporting only the published reliability estimates, if at all) or separate estimates by language group. Large-scale studies

ZHUHPRUHOLNHO\WRLQFOXGH VDPSOHVSHFL¿FHVWLPDWHV DQGKDGVXI¿FLHQWVDPSOH

sizes to report estimates by subgroups. When reported, the estimates of internal

FRQVLVWHQF\LQWKHVWXGLHVUHYLHZHGZHUHJHQHUDOO\ID-

YRUDEOHWKRXJKQRWDOZD\VZLWKLQDFFHSWDEOHUDQJHVIRU

Spanish assessments. For example, the reported reliability estimates for the Spanish version of the Story and Print Concepts used in the FACES 2000 and FACES

VWXGLHVZHUHPXFKORZHUWKDQWKHUHOLDELOLW\HVWL- mates for the English version.

Different factors can affect the reliability of measures

LQFOXGLQJKRZWDUJHWHGWKHLWHPVDUHWRWKHFRQVWUXFW

being measured and to the sample of children taking the

DVVHVVPHQWKRZZHOOLWHPVGLVFULPLQDWHDPRQJFKLO-

GUHQDQGWKHQXPEHURILWHPV8VXDOO\PHDVXUHVZLWK PRUHLWHPVZLOOGHPRQVWUDWHJUHDWHUUHOLDELOLW\WKRXJK ZHOOFRQVWUXFWHGDGDSWLYHDVVHVVPHQWVFDQDWWDLQKLJKO\

Table 1.Common Types of Reliability and Validity Evidence

Term 'HÀQLWLRQ

Internal consistency reliability

Indicates whether the items within a measure are measuring the same underlying concept

Test–retest reliability Indicates whether assessment of a particular construct (what is being measured) would result in the same score if repeat- ed later

Differential item functioning Notes whether the items within a measure function in the same way for different groups of children

Validity Indicates whether an assessment is measuring what it pur- ports to measure and under what conditions;; assessments can be valid for some uses and not for others

Predictive validity Indicates whether the assessment is related in expected ways to similar or related outcomes measured at a later time

(3)

UHOLDEOHHVWLPDWHVZLWKDOLPLWHGQXPEHURILWHPV0DQ\ RIWKHDVVHVVPHQWVZHUHDGDSWLYHDVVHVVPHQWVZLWK

items targeted to the children’s developmental levels.

+RZHYHUZKHQFRPSDUHGZLWKDQ(QJOLVKGRPLQDQW

sample, there may be greater differences in experiences, languages, and dialects among DLLs that affect their

UHVSRQVHVWRLWHPVLQZD\VXQUHODWHGWRZKDWLVEHLQJ

measured. In those cases, more items may be needed

WRREWDLQUHOLDEOHHVWLPDWHV7KRXJKRIWHQZHDNHUIRU

Spanish assessments, reported internal consistency

HVWLPDWHVZHUHXVXDOO\ZLWKLQDFFHSWDEOHUDQJHVZKHQ DJUHDWHUQXPEHURILWHPVZDVDGPLQLVWHUHG7KH ZHDNHUHVWLPDWHVRIUHOLDELOLW\RQVRPH6SDQLVKDV- sessments relative to the English assessments may be due in part to the number of items administered. For example, in FACES 2009, the reliability estimates for

WKH6SDQLVKOLWHUDF\DVVHVVPHQWVLQWKHIDOOZHUHZHDN OHVVWKDQEXWWKH\ZHUHEDVHGRQDQDYHUDJHRI

only 14 administered items, compared to the average of 17–26 items administered to the children (including

VRPH'//VWDNLQJWKHDVVHVVPHQWVLQ(QJOLVKZKLFK

had stronger internal consistency estimates (about .80). Because many assessments used in early childhood research are adaptive measures—that is, they present

LWHPVRILQFUHDVLQJGLI¿FXOW\XQWLOWKHFKLOGLQFRUUHFWO\ UHVSRQGVWRDVSHFL¿HGQXPEHURILWHPV²WKHQXPEHU

of items administered differs according to the child’s responses and the stop rules of the assessment (this is

HODERUDWHGRQEHORZ0RUHUHVHDUFKLVQHHGHGVSHFL¿- cally about the validity of stop rules themselves for dif-

IHULQJSRSXODWLRQV²LQSDUWLFXODUKRZUHOLDEOHDQGYDOLG WKH\DUHZKHQDGPLQLVWHULQJDVVHVVPHQWVWR'//V

Stability

The stability of a score, or test–retest reliability, indi-

FDWHVZKHWKHUDQDVVHVVPHQWRIDSDUWLFXODUFRQVWUXFW ZRXOGUHVXOWLQWKHVDPHVFRUHLIUHSHDWHGDZHHNODWHU

or, for more stable constructs, months or years later.

$PRQJWKHVWXGLHVZHUHYLHZHGWHVW±UHWHVWUHOLDELOLW\ ZLWKLQDIRXUZHHNWLPHSHULRGZDVUHSRUWHGIRURQO\

one bilingual measure, the Bilingual English Spanish Oral Language Screener (BESOS; Peña, Bedore, Gutier-

UH]&OHOOHQ,JOHVLDV *ROGVWHLQLQSUHSDUDWLRQZLWK

slightly stronger reliability indicated for the Spanish than for the English version.

Evidence of stability across longer periods of time

PRUHWKDQWKUHHPRQWKVZDVDYDLODEOHIRUVHYHUDO PHDVXUHVRIYRFDEXODU\DQGOLWHUDF\LQWKHUHYLHZHG

studies, but these may not be a good measure of stabil-

LW\SDUWLFXODUO\GXULQJWKHSUHVFKRRO\HDUVZKHQ'//V

are likely to have increased exposure to English via educational opportunities. For some children, preschool

PD\EHWKH¿UVWHGXFDWLRQDOH[SHULHQFHDQGWKH¿UVW

exposure for some DLLs to English. This makes it very

GLI¿FXOWWRGLVHQWDQJOHGLIIHUHQWLDWHWKHVWDELOLW\RIWKH

assessment from the responsiveness of different children to the educational opportunities and the sensitiv-

LW\RIWKHPHDVXUHWRLQWHUYHQWLRQ&KLOGUHQZKRDUH OHDUQLQJ(QJOLVKIRUWKH¿UVWWLPHPLJKWEHH[SHFWHG WRVKRZJUHDWHUFKDQJHLQWKHLUNQRZOHGJHRI(QJOLVK ZRUGVWKDQFKLOGUHQZKRKDYHEHHQOHDUQLQJLWIRUDQ

extended period of time. Thus, the stability estimates

RYHUDORQJSHULRGRIWLPHZRXOGEHORZHUWKDQRYHUD WZRZHHNSHULRGSDUWLFXODUO\ZKHQVDPSOHVKDYHERWK

simultaneous and sequential DLLs. With that caveat in mind, examination of correlations suggested that literacy measures might be less stable than vocabulary

PHDVXUHV'LFNLQVRQ0F&DEH&ODUN&KLDUHOOL :ROI +DPPHU'DYLVRQ/DZUHQFH 0LFFLR DOWKRXJKWKLVZDVQRWDOZD\VWKHFDVH$QWKRQ\HWDO +RZHYHUWKHVDPSOHVDQGWKHOHQJWKRIWLPH EHWZHHQDVVHVVPHQWVYDULHGPDNLQJLWPRUHGLI¿FXOWWR GUDZYDOLGLQIHUHQFHVIURPWKHHVWLPDWHV

How fair is the assessment?

Bias from a number of sources may render an assessment unfair or invalid for a particular group of children.

7KHTXHVWLRQVRUSUREHVPD\XVHIRUPDWVRUZRUGVWKDW

are unfamiliar in some cultures or the questions may

VDPSOHNQRZOHGJHWKDWLVVSHFL¿FWRDFHUWDLQFXOWXUH

or linguistic group. Under these conditions, the items

ZRXOGQRWEHUHSUHVHQWDWLYHRIWKHH[SHULHQFHVDQG NQRZOHGJHRIVRPHRIWKHFKLOGUHQDVVHVVHG,QVKRUW

the evidence suggests that some approaches to assess- ing young DLLs may not result in a fair representation

RIWKHFKLOGUHQ¶VNQRZOHGJH)RUH[DPSOHDVVHVVLQJWKH FRQFHSWXDOYRFDEXODU\RID'//²WKDWLVZKHWKHUDFKLOG KDVZRUGVIRUGLIIHUHQWREMHFWVDFWLRQVDQGFRQFHSWV²

(4)

ZRUGVWKHFKLOGNQRZV&KLOGUHQW\SLFDOO\DFTXLUHZRUGV IRUREMHFWVDQGDFWLYLWLHVH[SHULHQFHGDWKRPHLQWKHODQ-

JXDJHXVHGPRVWRIWHQDWKRPHDQGZRUGVIRUDFDGHPLF

concepts in the language used in school (Bialystok et al. 2010). Using one of the most commonly applied measures of English vocabulary (PPVT-III), Bialystok and colleagues (2010) noted that test items referring

WRKRPHREMHFWVDQGDFWLYLWLHVZHUHPRUHGLI¿FXOWIRU

Spanish-dominant children than for English-dominant

FKLOGUHQZKLOHVFKRROUHODWHGZRUGVSRVHGVLPLODUGLI-

¿FXOW\DFURVVJURXSV

Basal and Ceiling Rules

Implications of this phenomenon may go beyond the

FRUUHFWQHVVRILQGLYLGXDOLWHPVZKHQDVVHVVPHQWVRI

young children are adaptive, using basal and ceiling (or start and stop) rules. The items in adaptive assessments, as mentioned earlier, are generally ordered in

WHUPVRIGLI¿FXOW\EDVHGRQWKHUHVSRQVHVRIWKHQRUPD- tive sample, and the stop rules are designed so children

ZLOOKDYHDYHU\ORZSUREDELOLW\RIJHWWLQJDQ\LWHPV

beyond the stopping point correct; the scoring assumes responses to all subsequent items are incorrect. Because

LWHPVFRQFHUQLQJWKHQDPHVRIREMHFWVIRXQGLQWKH

home are easier for young children to correctly name, they are presented early in assessments of vocabulary.

&KLOGUHQZKRDUH'//VKRZHYHUPD\NQRZWKH(QJOLVK QDPHVIRUREMHFWVDQGDFWLYLWLHVUHODWHGWRVFKRRODQG DFDGHPLFVEXWQRWIRUWKRVHDVVRFLDWHGZLWKWKHKRPH

Using the published ceiling rules could, therefore, result in underestimates of their English vocabulary. None of

WKHUHYLHZHGVWXGLHVH[DPLQHGWKHDSSURSULDWHQHVVRI

basal and ceiling rules for young DLLs. 'LIIHUHQWLDO,WHP'LIÀFXOW\

Greater attention is needed for evaluating the fairness of individual items and tasks across subgroups. Some tasks may not have the same meaning in development

DFURVVVXEJURXSVRUWKHGLI¿FXOW\PD\YDU\DFURVV

languages. For example, rhyming is usually easier for

\RXQJFKLOGUHQZKHQWKHZRUGVKDYHRQO\RQHV\OODEOH

but English has many more one-syllable rhymes than

6SDQLVK/LPLWHGHYLGHQFHZDVSURYLGHGLQWKHUHYLHZHG

studies for the congruence of estimates of item- dif-

¿FXOW\DFURVVODQJXDJHVWKRXJKPDQ\RIWKHUHYLHZHG

UHVHDUFKVWXGLHVLQFOXGHGFKLOGUHQZKRVSRNHSULPDULO\ (QJOLVKDQGFKLOGUHQZKRVSRNHSULPDULO\6SDQLVKRU

another language. Although 74 percent of the studies assessed children in both languages, only one (Bialys- tok, Luk, Peets, & Yang, 2010) reported examination of

KRZLWHPVIXQFWLRQHGIRUGLIIHUHQWJURXSV3DUWLFXODUO\ ZKHQUHVHDUFKHUVWUDQVODWHLWHPVRUGHYHORSWKHLURZQ

assessments, attention is needed to assure that the assessments are fair representations of children’s skills,

NQRZOHGJHDQGDELOLW\7HVWGHYHORSHUVDQGUHVHDUFKHUV

should evaluate the items in tests to determine if they have the same meaning across groups—that is, indicat-

LQJZKHWKHUWKHLWHPVIXQFWLRQLQWKHVDPHZD\IRUGLI- ferent groups of children.

Parent and Teacher Reports

Some studies, particularly for children younger than

WZR\HDUVXVHGDVVHVVPHQWVEDVHGRQSDUHQWRUWHDFKHU

reports. Early childhood researchers often rely more on

WHDFKHUUHSRUWVWKDQSDUHQWUHSRUWVGXHWRHI¿FLHQFLHV LQGDWDFROOHFWLRQDQGSRWHQWLDOV\VWHPDWLFELDVZLWKSDU-

HQWUHSRUWLQJ3DQ5RZH6SLHU7DPLV/H0RQGD

Rescorla, Ratner, & Jusczyk, & Jusczyk, 2005; Reese & Read, 2000; Roberts Burchinal, & Durham, 1999). With English monolingual samples, correlations of teacher

UHSRUWZLWKGLUHFWDVVHVVPHQWVUDQJHIURPPRGHUDWHWR

high during the elementary school years, but suggest

SRWHQWLDOELDVHVZLWKWHDFKHUUHSRUWVSDUWLFXODUO\IRU LWHPVWKDWKDYHKLJKHULQIHUHQFHOHYHOV3HUU\ 0HLVHOV

1996). Less research is available for the infant-toddler

\HDUVDQGWKHUHVHDUFKZLWK'//VLVSDUWLFXODUO\OLPLWHG LQWKLVUHJDUG+RZHYHUDVWXG\E\9DJKDQGFROOHDJXHV 9DJK3DQ 0DQFLOOD0DUWLQH]VXJJHVWVWKDW

parents (rather than teachers) may provide the most valid information about DLLs’ vocabulary development.

$PRQJ'//VVWURQJHUUHODWLRQVKLSVZHUHIRXQGEH-

WZHHQSDUHQWUHSRUWVRIFKLOGUHQ¶VYRFDEXODU\DQGGLUHFW DVVHVVPHQWRIYRFDEXODU\WKDQEHWZHHQWHDFKHUUHSRUWV

of those children’s vocabulary and direct assessment of

WKHLUVNLOOV9DJK3DQ 0DQFLOOD0DUWLQH]

Conceptual Scoring

Other studies used measures that are conceptually

VFRUHG&RQFHSWXDOVFRULQJDOORZVSUREHVRUUHVSRQVHV LQPRUHWKDQRQHODQJXDJHDQGZLOOXVXDOO\SUHVHQW

(5)

DPRUHYDOLGDVVHVVPHQWZKHQPHDVXULQJFKLOGUHQ¶V NQRZOHGJHRIFRQFHSWVUDWKHUWKDQYRFDEXODU\LQD SDUWLFXODUODQJXDJH)RUUHFHSWLYHYRFDEXODU\KRZ- ever, comparisons across linguistic groups can be more

FKDOOHQJLQJ,I\RXSUHVHQWIRXUSLFWXUHVRIREMHFWVDQG

the name of one of them in one language (giving a 1 in 4 chance of selecting the correct picture) and the child

LVXQVXFFHVVIXOWKHQIROORZZLWKDSURPSWLQWKHRWKHU ODQJXDJHQRZJLYLQJDLQFKDQFHRIVXFFHVVWKH SUREDELOLW\RIVXFFHVVIDYRUVFKLOGUHQZKRNQRZWZR ODQJXDJHVRYHUFKLOGUHQZKRNQRZRQO\RQH7KLVLVQRW

a problem for expressive vocabulary because a child

FRXOGGUDZXSRQWKRXVDQGVRIGLIIHUHQWZRUGVWRQDPH

a picture.

It is important that researchers consider all sources of

SRWHQWLDOELDVZKHQSUHVHQWLQJDQGLQWHUSUHWLQJWKHLU ¿QGLQJVDQGZKHQIHDVLEOHFRQWULEXWHWRWKHNQRZO-

HGJHEDVHDERXWYDOLGLW\RIPHDVXUHVZLWK'//V

Does the Assessment

Answer the Questions of Interest?

If assessments of DLLs are to be valid for the goals of the research, care must be taken to select instruments

DQGPHWKRGVWKDWPDWFKWKHTXHVWLRQWREHDQVZHUHG )RULQVWDQFHLWLVLPSRUWDQWWRFRQVLGHUZKHWKHUWKHUH-

VHDUFKTXHVWLRQUHTXLUHVPHDVXUHPHQWRIVNLOOVNQRZO- edge, or behavior at a single point in time or if the focus

LVRQFKDQJHSDUWLFXODUO\LIWKHFKDQJHPD\UHÀHFWWKH

effectiveness of a program or intervention. In the latter case, researchers should consider the language used in the program or intervention and the possibility that they

ZLOOQHHGWRDVVHVVWKHFRQVWUXFWVRILQWHUHVWDWEDVHOLQH

in both the child’s home language and the language used in the program to have a valid estimate of the program’s effects. Without an assessment in the home language,

FKLOGUHQPD\DSSHDUWRODFNVNLOOVDWEDVHOLQHDQGVKRZ JDLQVLQPDQ\VNLOOVZKHQDVVHVVHGDWDODWHUSRLQWLQ WLPHZKHQWKHRQO\NQRZOHGJHWKDWWKH\DFTXLUHGZDV

comprehension of the questions in English. A program

ZLWKPDQ\'//VFRXOGDSSHDUWREHPDNLQJJUHDWHU JDLQVWKDQDSURJUDPLQZKLFKFKLOGUHQHQWHUZLWKRXW

any skills at all. For example, if the children already pos- sess understanding of a concept such as seriation (for example, tall, taller, and tallest), but cannot understand

WKHTXHVWLRQWKH\ZRXOGGRSRRUO\RQWKRVHLWHPVDW EDVHOLQHEXWGRZHOORQFHWKH\XQGHUVWDQGWKHTXHVWLRQ LQ(QJOLVK7KHFKDQJHZDVLQWKHLUXQGHUVWDQGLQJRI

English rather than their understanding of seriation. In

DQRWKHUSURJUDPFKLOGUHQPD\HQWHUZLWKRXWXQGHU-

VWDQGLQJRIVHULDWLRQDQGVRZRXOGQHHGWROHDUQERWK WKH(QJOLVKDQGVHULDWLRQ$WWKHHQGRIWKH\HDUZKHQ

children in both groups successfully responded to the seriation tasks, the program that taught the children the English, but did not extend mathematical thinking,

ZRXOGORRNDVWKRXJKLWKDGDJUHDWHUHIIHFWRQFRJQLWLYH

development than it actually did.

What Levels of Performance and

Progress Are Expected for DLLs

on Different Assessments?

0RUHLQIRUPDWLRQLVQHHGHGDERXWH[SHFWHGSHUIRU- mance of DLLs on assessments. To provide information about their expected performance on language measures, developers should provide supplemental norms for DLLs or estimates of the mean and standard devia-

WLRQIRUWKHVXEVDPSOHRI'//V0DQ\VDPSOHVRI'//V LQWKHVWXGLHVZHUHYLHZHGZHUHUHODWLYHO\VPDOODQG

often the only norms the researchers had available for

LQWHUSUHWLQJWKHVFRUHVZHUHEDVHGRQVWDQGDUGL]DWLRQ VDPSOHVLQZKLFKPRVWRIWKHFKLOGUHQZHUHPRQROLQ- gual in either English or Spanish.

0RVWUHVHDUFKVWXGLHVFRQGXFWHGZLWK'//VLQWKH

United States in the last 10 years included primarily or

RQO\ORZLQFRPH6SDQLVK±(QJOLVK'//VOLPLWLQJWKH JHQHUDOL]DELOLW\RIWKH¿QGLQJVWRRWKHUJURXSVRI'//V 1DWLRQDOO\KRZHYHUWKHPDMRULW\RI\RXQJ'//VUHVLGH LQKRPHVZLWKOLPLWHGLQFRPHDQGKDYH6SDQLVKDVD

home language (Shin and Kominski, 2010).

,WZRXOGEHKHOSIXOLIUHVHDUFKHUVSDUWLFXODUO\IRUODUJH

scale studies) provided separate information about the mean scores and study characteristics of their DLL samples (including socioeconomic background, range of dialects represented, and age range) and their skills.

(6)

Vocabulary Assessments Used with DLLs in

Reviewed Studies

Assessment Abrieviations Title, Author, Date

English Vocabulary

CDI MacArthur-Bates Communicative Development Inventories (Fenson et al., 1993)

EOWPVT Expressive One-Word Picture Vocabulary Test (Brownell, 2000a) ROWPVT Receptive One-Word Picture Vocabulary Test (Brownell, 2000b) PPVT

PPVT–R Peabody Picture Vocabulary Test–Revised (Dunn & Dunn, 1981) PPVT–III Peabody Picture Vocabulary Test–III (Dunn & Dunn, 1997) PPVT–4 Peabody Picture Vocabulary Test–4 (Dunn & Dunn, 2007

WJ–III (Picture Vocabulary) :RRGFRFN/DQJXDJH3UR¿FLHQF\%DWWHU\±5HYLVHG(QJOLVK)RUP:RRGFRFN 1995)

Spanish Vocabulary

Inventario (CDI Spaish Edition) El Inventario del Desarrollo de Habilidades Comunicativas (Jackson-Maldona- do, Thal, Marchman, Newton, Fenson, & Conboy, 2003)

TVIP Test de Vocabulario en Imágenes Peabody (Dunn, Padilla, Lugo, & Dunn, 1986) WM–II (Vocabulario Sobre

Dibujos)

:RRGFRFN/DQJXDJH3UR¿FLHQF\%DWWHU\±5HYLVHG6SDQLVK)RUP:RRGFRFN Muñoz-Sandoval, 1995)

Conceptuall Scored (English & Spanish)

EOWPVT-SBE Expressive One-Word Picture Vocabulary Test: Spanish Bilingual Edition _{(Brownell, 2001a)} ROWPVT-SBE Receptive One-Word Picture Vocabulary Test : Spanish Bilingual Edition _{(Brownell, 2001b)} SEVC Spanish–English Vocabulary Checklist (Patterson, 1998)

Chiese Vocabulary

PPVT–R Chinese Peabody Picture Vocabulary Test—Revised: Chinese (Lu & Liu, 1998)

What Does the Score Mean and

How Can We Interpret the Results?

6WXG\UHVXOWVVKRXOGEHLQWHUSUHWHGDQGGLVFXVVHGZLWK

reference to the characteristics of the children in the study. When comparing differences in the performance and progress of groups of children, information about differences in opportunity to learn (due to different socioeconomic, cultural, or linguistic backgrounds) should be included. When discussing standard scores, researchers should help readers understand the similarities and

GLIIHUHQFHVEHWZHHQWKHQRUPDWLYHJURXSDQGWKHVWXG\

sample. Even for the same sample of children, mean standard scores vary across different measures of the same construct.

The baseline information in FACES 2009 (Aikens et al., 2011), as one example, provides potentially different

LQWHUSUHWDWLRQVRI'//V¶NQRZOHGJHRIYRFDEXODU\DQG

concepts depending on the dimensions assessed and the norm group used for deriving standard scores. FACES 2009 assessed vocabulary—the area most commonly assessed among young DLLs (see Table 2)—using multiple measures and provided standard scores for 4-year-old DLLs based on the PPVT-4 (receptive English vocabulary), the English version of the EOWPVT (expressive vocabulary), and the Spanish bilingual version (EOW- PVT-SBE), in addition to scores on the TVIP, a Spanish receptive vocabulary measure. Although mean scores

(7)

IRUWKHSUHGRPLQDQWO\(QJOLVKVSHDNLQJ'//VZHUHLQD

similar range across the English assessments, the scores

IRUWKH6SDQLVKVSHDNLQJ'//VYDULHGPRUHZLGHO\RQWKH (QJOLVKDQGELOLQJXDODVVHVVPHQWVZLWKPHDQVWDQGDUG

scores of 56, 67, and 86 on the PPVT-4, and the English and bilingual norms for the EOWPVT-SBE, respectively.

6WDQGDUGVFRUHVIRU6SDQLVKVSHDNLQJ'//VZHUHPRUH VLPLODUEHWZHHQWKH79,3DQGWKH(2:3976%( DQGUHVSHFWLYHO\7KHVHODWWHUWZRPHDVXUHVKDYH

standard scores based on Spanish-speaking or bilingual

VDPSOHVRIFKLOGUHQZLWKKDOIRUPRUHRIWKHVDPSOHIURP KRPHVZLWKOLPLWHGPDWHUQDOHGXFDWLRQ7KHVHGLIIHUHQF-

HVDQGVLPLODULWLHVKLJKOLJKWWKHGLI¿FXOW\LQLQWHUSUHWLQJ FKLOGUHQ¶VVNLOOVZLWKRXWLQIRUPDWLRQDERXWWKHDVVHVV- ments and the normative samples used to generate the standard scores.

Do the Assessments Measure What They

Are Supposed to Measure?

The studies provided some additional evidence of valid-

LW\²WKDWLVWKDWWKHDVVHVVPHQWVPHDVXUHGZKDWWKH\ ZHUHVXSSRVHGWRPHDVXUH7KHW\SHRIHYLGHQFHPRVW IUHTXHQWO\LGHQWL¿HGIRUODQJXDJHDQGOLWHUDF\DVVHVV-

PHQWVRI'//VZDVDQDVVRFLDWLRQZLWKFKLOGUHQ¶VDJH

or exposure to English, including moderate correlations

RIYRFDEXODU\DQGODQJXDJHDVVHVVPHQWVZLWKDJHDQG SDUHQWUHSRUWHGH[SRVXUHWR(QJOLVKDVZHOODVHYLGHQFH

of an increase in assessment scores across time.

6LPLODUWRUHOLDELOLW\FRHI¿FLHQWVWKHUDQJHRIUHSRUWHG FRQFXUUHQWYDOLGLW\FRHI¿FLHQWVZDVZHDNHUZKHQFRP-

SDUHGWRFRHI¿FLHQWVIRXQGDFURVVPHDVXUHVLQVWXGLHVRI

young monolingual English samples.4_{They ranged from}

ORZPRGHUDWHWRVWURQJZLWKVWURQJHUHVWLPDWHVEHWZHHQ YRFDEXODU\DQGODQJXDJHPHDVXUHVDQGZHDNHUHVWLPDWHV EHWZHHQYRFDEXODU\DQGOLWHUDF\PHDVXUHV

Do the Assessments Support Understanding

of the School Readiness of DLLs?

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

concurrent or predictive validity separately for monolingual and DLL children. The sample sizes of the peer-re-

YLHZHGUHVHDUFKVWXGLHVZHUHRIWHQWRRVPDOOIRUVHSDUDWH

subgroup analyses of validity. Other than the large-scale government studies, only three of the research studies had sample sizes greater than 200. Predictive validity

HYLGHQFHLQWKHVHVWXGLHVXVXDOO\H[DPLQHGZKHWKHUFKLO-

GUHQ¶VVFRUHVLQFUHDVHGDFURVVWLPHSRLQWVDQGZKHWKHU

earlier vocabulary and language assessments predicted later literacy.

The studies that provided information about the predictive validity of the measures in relation to school readi-

QHVVGLGQRWH[DPLQH¿QGLQJVIRUPHDVXUHVVHSDUDWHO\E\

subgroup. The report for FACES 1997 (Zill et al., 2003)

RQO\LQFOXGHGFKLOGUHQZKRWRRNWKH(QJOLVKDVVHVVPHQW DWHDFKWLPHSRLQWDQGGLGQRWFOHDUO\LQGLFDWHKRZPDQ\

of them had Spanish as a home language. One study of DLLs reported analysis of predictive relationships from

SUHVFKRROWR¿UVWJUDGHUHDGLQJLQ(QJOLVKZLWKWKHIXOO

sample and did not estimate differences by language

SUR¿FLHQF\5LQDOGL 3iH]:LWKWKHIXOOVDPSOH WKDWVWXG\LQGLFDWHGZHDNUHODWLRQVKLSVIRULQGLYLGXDO

subtests, although a combination of several subtests across both Spanish and English increased the amount of explained variance in English reading. Hammer and colleagues (2007) used a combination of language measures

WRH[DPLQHWKHUHODWLRQVKLSEHWZHHQSUHVFKRROODQJXDJH

development and spring kindergarten reading, compar-

LQJWKHVNLOOVRI'//VZKROHDUQHG(QJOLVKDWKRPHSULRU WRVWDUWLQJSUHVFKRRO+HDG6WDUWWRWKRVHRI'//VZKR ZHUHLQWURGXFHGWR(QJOLVKDW+HDG6WDUW$OWKRXJKWKH

analyses provided estimates for each of the subgroups, the evidence cannot be attributed to a single measure, but rather to the component measures.5

0DQ\IDFWRUVFDQDIIHFWWKHVWUHQJWKRIYDOLGLW\FRHI-

(8)

UHODWLRQVKLSEHWZHHQDVVHVVPHQWVZLOOGHSHQGRQPDQ\

different factors, including the reliability of the assessment, similarity in mode of assessment, similarity in

WKHGLPHQVLRQVDQGFRQVWUXFWVDVVHVVHGWLPHEHWZHHQ

assessments, and age of the child (the younger the child

DW¿UVWDVVHVVPHQWWKHZHDNHUWKHUHODWLRQVKLS7KH

predictive validity evidence of early childhood assessments among English monolingual samples typically

GHPRQVWUDWHVORZWRPRGHUDWHSUHGLFWLYHFRUUHODWLRQ FRHI¿FLHQWVZLWKOHVVWKDQSHUFHQWRIWKHRYHUDOOYDUL- ance in early academic performance predicted from any single preschool measure (LaParo & Pianta, 2000). When samples include DLLs, the number of additional variables

WKDWFDQDIIHFWWKHVWUHQJWKRIWKHFRHI¿FLHQWLQFUHDVHV²

for example, the age of introduction to the language used, the amount of exposure to the language of assessment, and intervention or preschool experiences. Of the stud-

LHVUHYLHZHGYRFDEXODU\ZDVDVVHVVHGPRVWRIWHQ\HW HYLGHQFHRIDUHODWLRQVKLSWRVFKRRORXWFRPHVZDVIRXQG RQO\ZLWKODWHQWWUDLWVFRPELQLQJPXOWLSOHDVSHFWVRI ODQJXDJH+DPPHU/DZUHQFHDQG0LFFLR5LQDOGL DQG3iH]0RUHUHVHDUFKLVQHHGHGWRGHWHUPLQH KRZWRPRQLWRUWKHGHYHORSPHQWRI'//VWRHQVXUHODWHU

success in school.

Conclusions

0RUHLQIRUPDWLRQLVQHHGHGDERXWRXUFXUUHQWDVVHVV-

PHQWVDQGZKDWLQIHUHQFHVFDQEHGUDZQIURPWKHLUUH-

VXOWV$OVRQHHGHGDUHDVVHVVPHQWVIRUXVHZLWK'//VZKR

speak languages other than English. Researchers should

FRQVLGHUFDUHIXOO\ZKDWDVVHVVPHQWZLOODQVZHUWKHTXHV- tions of interest and report more information about the samples and measures.

Researchers need to consider if assessments are valid for the children in their samples. When samples include children from multiple linguistic back-

JURXQGVZLOOWKHPHWKRGVDQGLWHPVIDLUO\UHSUHVHQW DOOFKLOGUHQ¶VNQRZOHGJHVNLOOVDQGEHKDYLRUV"$UH

cultural or linguistic biases inherent in the use of the

DVVHVVPHQWZLWKDSDUWLFXODUJURXSRIFKLOGUHQ"(YHQ ZLWKLQDVLQJOHOLQJXLVWLFJURXSGLIIHUHQFHVLQGLDOHFW

may bias results unless the assessment accounts for them.

5HVHDUFKHUVVKRXOGFRQVLGHUZKHWKHUWKHVHOHFWHG DVVHVVPHQWVDUHELDVHGLQDQ\ZD\'RWKHWDVNVRU

items require similar levels of skill across languages

DQGFXOWXUHV",VWKHWDVNHTXDOO\UHSUHVHQWDWLYHRI VNLOOVDFURVVGLIIHUHQWJURXSV"

7HVWGHYHORSHUVLQFOXGLQJUHVHDUFKHUVZKRGHYHORS

forms in other languages) should provide evidence that the items are equivalent for children from different groups, that is, that the items contribute to mea-

VXUHPHQWRIWKHFRQVWUXFWLQWKHVDPHZD\VDFURVV JURXSVDQGWKDWWKHGLI¿FXOW\RIHDFKRIWKHLWHPVRU

tasks is the similar across groups.

0RUHSUHGLFWLYHYDOLGLW\HYLGHQFHLVQHHGHGSDUWLFX- larly for infant-toddler measures. While this is true for early childhood assessments in general, measures and evidence of predictive validity are particularly sparse for young DLLs.

(9)

(Endnotes)

1. 7KHUHYLHZLQFOXGHGSHHUUHYLHZHGMRXUQDODUWLFOHVSXE-

OLVKHGEHWZHHQDQGZLWKVDPSOHVIURPWKH United States or its territories and Canada that included at least one direct child assessment or standardized rating of the language or literacy development of DLL children prior to age 6 or kindergarten entry. To examine psycho- PHWULFSURSHUWLHVZHH[FOXGHGVWXGLHVZLWKOHVVWKDQ DLLs and those that only analyzed language samples, UHVXOWLQJLQSHHUUHYLHZHGUHVHDUFKDUWLFOHV6SDQLVK DQG(QJOLVKZHUHWKHPRVWFRPPRQO\UHSRUWHGODQJXDJHV in the samples, and more studies focused on preschoolers

WKDQRQLQIDQWVDQGWRGGOHUV:HDOVRUHYLHZHGJRYHUQ-

ment reports of large-scale studies of early childhood that included at least one direct child assessment of language DQGRUOLWHUDF\$PRQJJRYHUQPHQWUHSRUWVSXEOLVKHG LQWKHODVW\HDUVZHORFDWHGRQO\VHYHQODUJHVFDOH national studies that examined children’s language or literacy development prior to kindergarten entry and included DLLs in any of these assessments. Description of the sample characteristics and study purpose for each of the studies can be found in the full report (Bandel et DOZKLFKLVDYDLODEOHRQWKH&(&(5'//ZHEVLWH >KWWSFHFHUGOOISJXQFHGX@

2. Researchers and assessment developers often require that assessment tools have evidence of reliability values of 0.70 or higher to support inferences about the measure %DFRQ&RKHQ/LWZLQ1XQQDOO\ KRZHYHUWKHPLQLPDOUHFRPPHQGHGOHYHORILQWHUQDO consistency differs according to the type of inference that ZLOOEHPDGHDERXWWKHUHVXOWV

3. The reported reliability estimates for the Spanish version of the Story and Print Concepts used in the FACES 2000

DQG)$&(6VWXGLHVZHUHOHVVWKDQZKLOHUHOL-

DELOLW\HVWLPDWHVIRUWKH(QJOLVKYHUVLRQZHUHJUHDWHUWKDQ .70.

4. &RUUHODWLRQVZLWKWKH'//VDPSOHVUDQJHGIURPWR

.79. Among English-only samples, estimates are often stronger. Some examples of correlations among English DVVHVVPHQWVLQFOXGHWKH3/6DQGWKH&(/)3ZLWKU IRUWRWDOVFRUHVU IRUH[SUHVVLYHODQJXDJHVFRUHV from each of these assessments. Scores on the PPVT-4 UHFHSWLYHYRFDEXODU\ZLWKWKH([SUHVVLYH9RFDEXODU\ Test-Second Edition (EVT-2) ranged from .80 to .84.

5. 7KH(QJOLVKODQJXDJHFRPSRQHQWPHDVXUHZDVEDVHGRQ

the PPVT-III and the Receptive Language subtest of the 7(/'7KH6SDQLVKODQJXDJHFRPSRQHQWPHDVXUHZDV based on the TVIP and the Auditory Comprehension sub- WHVWRIWKH3/67KHVWXG\IRXQGWKDWWKHJURZWKRQWKH component measure during the preschool year predicted literacy scores at the end of kindergarten.

References

$LNHQV1+XOVH\/.0RLGXGGLQ(.RSDFN$7DN\L Laryea, A., Tarullo, L., & West, J. (2011). Data tables for FACES 2009 Head Start children, families, and pro- JUDPV3UHVHQWDQGSDVWGDWDIURP)$&(6UHSRUW$&) OPRE 2011-33b). U.S. Department of Health and Human 6HUYLFHV2I¿FHRI3ODQQLQJ5HVHDUFKDQG(YDOXDWLRQ Administration for Children and Families. Washington, '&86*RYHUQPHQW3ULQWLQJ2I¿FH $QWKRQ\-/6RODUL(-:LOOLDPV-06FKRJHU.' =KDQJ=%UDQXP0DUWLQ/ )UDQFLV'-'HYHORSPHQWRIELOLQJXDOSKRQRORJLFDODZDUHQHVVLQ 6SDQLVKVSHDNLQJ(QJOLVKODQJXDJHOHDUQHUV7KHUROHV RIYRFDEXODU\OHWWHUNQRZOHGJHDQGSULRUSKRQRORJLFDO DZDUHQHVV6FLHQWL¿F6WXGLHVRI5HDGLQJ(6), 535–564. $XG6+XVVDU:-RKQVRQ).HQD*5RWK(0DQ-

ning, E., … Zhang, J. (2012). 7KHFRQGLWLRQRIHGXFDWLRQ (NCES 2012-045). U.S. Department of Education, National Center for Education Statistics. Washington, '&86*RYHUQPHQW3ULQWLQJ2I¿FH

Bacon, D. (2004). The contributions of reliability and pretests to effective assessment. 3UDFWLFDO$VVHVVPHQW5HVHDUFK (YDOXDWLRQ(3).

Bandel, E., Atkins-Burnett, S., Castro, D. C., Wulsin, C. S., &

3XWQDP0([DPLQLQJWKHXVHRIODQJXDJHDQG

OLWHUDF\DVVHVVPHQWVZLWK\RXQJGXDOODQJXDJHOHDUQHUV Research report no. 1. Center for Early Care and Educa- tion Research–Dual Language Learners (CECER-DLL). &KDSHO+LOO8QLYHUVLW\RI1RUWK&DUROLQD)UDQN3RUWHU Graham Child Development Institute.

%DUQHWW:6&DURODQ0()LW]JHUDOG- 6TXLUHV-+ (2011). 7KHVWDWHRISUHVFKRRO6WDWHSUHVFKRRO\HDU-

ERRN1HZ%UXQVZLFN1-1DWLRQDO,QVWLWXWHIRU(DUO\

Education Research.

Bialystok, E., Luk, G., Peets, K. F., & Yang, S. (2010). Recep- tive vocabulary differences in monolingual and bilingual children. %LOLQJXDOLVP/DQJXDJHDQG&RJQLWLRQ(4), 525–531. %URZQHOO5D([SUHVVLYHRQHZRUGSLFWXUHYRFDEX- ODU\WHVW1RYDWR&$$FDGHPLF7KHUDS\3XEOLFDWLRQV %URZQHOO5E5HFHSWLYHRQHZRUGSLFWXUHYRFDEX- ODU\WHVW1RYDWR&$$FDGHPLF7KHUDS\3XEOLFDWLRQV %URZQHOO5D([SUHVVLYHRQHZRUGSLFWXUHYRFDEX-

lary test6SDQLVK%LOLQJXDO(GLWLRQ0DQXDO1RYDWR

&$$FDGHPLF7KHUDS\3XEOLFDWLRQV

%URZQHOO5E5HFHSWLYHRQHZRUGSLFWXUHYRFDEXODU\

test6SDQLVK%LOLQJXDO(GLWLRQ0DQXDO1RYDWR&$

(10)

Cohen, J. (1977).6WDWLVWLFDOSRZHUDQDO\VLVIRUWKHEHKDYLRUDO

sciencesUHYHG1HZ<RUN$FDGHPLF3UHVV

'LFNLQVRQ'.0F&DEH$&ODUN±&KLDUHOOL1 :ROI A. (2004). Cross-language transfer of phonological DZDUHQHVVLQORZLQFRPH6SDQLVKDQG(QJOLVKELOLQJXDO preschool children. $SSOLHG3V\FKROLQJXLVWLFV(03), 323–347.

Dunn, L., & Dunn, L. (1981). 3HDERG\SLFWXUHYRFDEXODU\ WHVW²UHYLVHG&LUFOH3LQHV01$PHULFDQ*XLGDQFH Service.

Dunn, L., & Dunn, L. (1997). 3HDERG\SLFWXUHYRFDEXODU\WHVW UGHG&LUFOH3LQHV01$PHULFDQ*XLGDQFH6HUYLFH Dunn, D., & Dunn, L. (2007). 3HDERG\SLFWXUHYRFDEXODU\WHVW

WKHG0LQQHDSROLV011&63HDUVRQ,QF

Dunn, L., Padilla, E. R., Lugo, D. E., & Dunn, L. (1986). Test

GHYRFDEXODULRHQLPDJHQHV3HDERG\&LUFOH3LQHV01 American Guidance Service.

Fenson, L., Dale, P., Reznick, S., Thal, D., Bates, E., Hartung, J., … Reilly, J. (1993). 0DF$UWKXUFRPPXQLFDWLYHGHYHO-

RSPHQWLQYHQWRULHV6DQ'LHJR&$6LQJXODU3XEOLVKLQJ +DPPHU&6/DZUHQFH)5 0LFFLR$:

Bilingual children’s language abilities and early reading outcomes in Head Start and kindergarten. /DQJXDJH 6SHHFKDQG+HDULQJ6HUYLFHVLQ6FKRROV(3), 237. +DPPHU&6'DYLVRQ0'/DZUHQFH)5 0LFFLR$

W. (2009). The effect of maternal language on bilingual children’s vocabulary and emergent literacy development during Head Start and kindergarten.6FLHQWL¿F6WXGLHVRI 5HDGLQJ(2), 99–121.

-DFNVRQ0DOGRQDGR'7KDO'-0DUFKPDQ91HZWRQ T., Fenson, L., & Conboy, B. (2003). (OLQYHQWDULRGHOGH-

VDUUROORGHKDELOLGDGHVFRPXQLFDWLYDV8VHU¶VJXLGHDQG WHFKQLFDOPDQXDO%DOWLPRUH0'%URRNHV3XEOLVKLQJ&R /D3DUR.0 3LDQWD5&3UHGLFWLQJFKLOGUHQ¶V FRPSHWHQFHLQWKHHDUO\VFKRRO\HDUV$PHWDDQDO\WLF UHYLHZ5HYLHZRI(GXFDWLRQDO5HVHDUFK(4), 443–484. /LWZLQ06+RZWRDVVHVVDQGLQWHUSUHWVXUYH\SV\- FKRPHWULFVQGHG7KRXVDQG2DNV&$6DJH3XEOLFD- tions, 2003.

Lu, L., & Liu, H. S. (1998). 7KH3HDERG\SLFWXUHYRFDEXODU\ WHVW²UHYLVHG&KLQHVH7DLSHL7DLZDQ3V\FKRORJLFDO Publishing. Nunnally, J. C. (1978). 3V\FKRPHWULFWKHRU\QGHG1HZ <RUN1<0F*UDZ+LOO 3DQ%$5RZH0/6SLHU( 7DPLV/HPRQGD& 0HDVXULQJSURGXFWLYHYRFDEXODU\RIWRGGOHUVLQ ORZLQFRPHIDPLOLHVFRQFXUUHQWDQGSUHGLFWLYHYDOLGLW\ of three sources of data. -RXUQDORI&KLOG/DQJXDJH 587-608.

Patterson, J. (1998). Expressive vocabulary development and ZRUGFRPELQDWLRQVRI6SDQLVK±(QJOLVKELOLQJXDOWRGGOHUV

$PHULFDQ-RXUQDORI6SHHFK±/DQJXDJH3DWKRORJ\(4), 46–56.

Peña, E., Bedore, L., Gutierrez-Clellen, V., Iglesias, A., & Goldstein, B. (in preparation). %LOLQJXDO(QJOLVK±6SDQLVK DVVHVVPHQW

3HUU\1( 0HLVHOV6-+RZDFFXUDWHDUHWHDFKHU

MXGJPHQWVRIVWXGHQWV¶DFDGHPLFSHUIRUPDQFH"(NCES :DVKLQJWRQ'&1DWLRQDO&HQWHUIRU(GXFDWLRQ Statistics. 5LQDOGL& 3iH]03UHVFKRROPDWWHUV3UHGLFWLQJ UHDGLQJGLI¿FXOWLHVIRU6SDQLVKVSHDNLQJELOLQJXDOVWX- GHQWVLQ¿UVWJUDGH/HDUQLQJ'LVDELOLWLHV(1), 71. 5HVFRUOD/5DWQHU1%-XVF]\N3 -XVF]\N$0 (2005). Concurrent validity of the language development

VXUYH\DVVRFLDWLRQVZLWKWKH0DF$UWKXU%DWHVFRPPX- QLFDWLYHGHYHORSPHQWLQYHQWRULHV:RUGVDQGVHQWHQFHV $PHULFDQ-RXUQDORI6SHHFKDQG/DQJXDJH3DWKRORJ\ (2), 156-163. 5HHVH( 5HDG63UHGLFWLYHYDOLGLW\RIWKH1HZ =HDODQG0DF$UWKXU&RPPXQLFDWLYH'HYHORSPHQW,QYHQ- WRUW\:RUGVDQG6HQWHQFHV-RXUQDORI&KLOG/DQJXDJH (2), 255-266. 5REHUWV-(%XUFKLQDO0 'XUKDP03DUHQWV¶ report of vocabulary and grammatical development of $IULFDQ$PHULFDQSUHVFKRROHUV&KLOGDQGHQYLURQPHQWDO associations. &KLOG'HYHORSPHQW92–106.

Shin, H. B., & Kominski, R. A. (2010). Language use in the

8QLWHG6WDWHVAmerican Community Survey 5HSRUWV$&686&HQVXV%XUHDX:DVKLQJWRQ'& 86*RYHUQPHQW3ULQWLQJ2I¿FH

9DJK6%3DQ%$ 0DQFLOOD0DUWLQH]-0HD-

VXULQJJURZWKLQELOLQJXDODQGPRQROLQJXDOFKLOGUHQ¶V HQJOLVKSURGXFWLYHYRFDEXODU\GHYHORSPHQW7KHXWLOLW\RI combining parent and teacher report.&KLOG'HYHORSPHQW (5), 1545-1563.

9RJHO&$%ROOHU.;XH<%ODLU5$LNHQV1%XUZLFN A., & Stein, J. (2011). /HDUQLQJDVZHJR$¿UVWVQDS-

VKRWRI(DUO\+HDG6WDUWSURJUDPVVWDIIIDPLOLHVDQG FKLOGUHQ (ACF-OPRE 2011-7). U.S. Department of Health DQG+XPDQ6HUYLFHV2I¿FHRI3ODQQLQJ5HVHDUFKDQG Evaluation, Administration for Children and Families. :DVKLQJWRQ'&86*RYHUQPHQW3ULQWLQJ2I¿FH

(11)

About CECER-DLL

&(&(5'//LVDQDWLRQDOFHQWHUWKDWLVEXLOGLQJFDSDFLW\IRUUHVHDUFKZLWKGXDOODQJXDJHOHDUQHUV'//VDJHVELUWK WKURXJK¿YH\HDUV&(&(5'//DLPVWRLPSURYHWKHVWDWHRINQRZOHGJHDQGPHDVXUHPHQWLQHDUO\FKLOGKRRGUHVHDUFK on DLLs, identify and advance research on best practices for early care and education programming, and develop and GLVVHPLQDWHSURGXFWVWRLPSURYHUHVHDUFKRQ'//V&(&(5'//LVDFRRSHUDWLYHDJUHHPHQWEHWZHHQWKH)UDQN3RUWHU *UDKDP)3*&KLOG'HYHORSPHQW,QVWLWXWHDW7KH8QLYHUVLW\RI1RUWK&DUROLQDDW&KDSHO+LOODQGWKH2I¿FHRI3ODQQLQJ 5HVHDUFK (YDOXDWLRQ235(LQWKH$GPLQLVWUDWLRQIRU&KLOGUHQ )DPLOLHV$&)LQFROODERUDWLRQZLWKWKH2I¿FHRI +HDG6WDUWDQGWKH2I¿FHRI&KLOG&DUH

Suggested citation

Atkins-Burnett, S., Bandel, E., & Aikens, N. (2012). 5HVHDUFKEULHI$VVHVVPHQWWRROVIRUWKHODQJXDJHDQGOLWHUDF\GH-

YHORSPHQWRI\RXQJGXDOODQJXDJHOHDUQHUV'//V&KDSHO+LOO7KH8QLYHUVLW\RI1RUWK&DUROLQD)3*&KLOG'HYHORSPHQW Institute, CECER-DLL. 7KLVEULHIVXPPDUL]HVUHVXOWVIURPDUHYLHZRIWKHOLWHUDWXUHVSRQVRUHGE\&(&(5'//FRQGXFWHGE\DUHVHDUFKWHDP FRQVLVWLQJRI(LOHHQ%DQGHO6DOO\$WNLQV%XUQHWW'LQD&&DVWUR&ODLUH6PLWKHU:XOVLQDQG0DULVD3XWQDPZLWK0DUJDUHW %XUFKLQDO/LVD/RSp]9HUD*XWLpUUH]&OHOOHQDQG(OOHQ3HLVQHU)HLQEHUJDVUHVHDUFKSDUWQHUV7KHZRUNZDVVXSSRUWHGE\ DFRRSHUDWLYHDJUHHPHQWIXQGHGE\WKH2I¿FHRI3ODQQLQJ5HVHDUFKDQG(YDOXDWLRQ235(86'HSDUWPHQWRI+HDOWK DQG+XPDQ6HUYLFHV3HUPLVVLRQWRFRS\GLVVHPLQDWHRURWKHUZLVHXVHLQIRUPDWLRQIURPWKLVGRFXPHQWIRUHGXFDWLRQDO purposes is granted, provided that appropriate credit is given.

$GGLWLRQDO5HVRXUFHV)RUDGGLWLRQDOLQIRUPDWLRQUHJDUGLQJWKLVUHVHDUFKEULHIVHHKWWSFHFHUGOOISJXQFHGX Woodcock, R. W. (1995). :RRGFRFNODQJXDJHSUR¿FLHQF\EDW- WHU\²UHYLVHG,WDVFD,/5LYHUVLGH3XEOLVKLQJ :RRGFRFN5: 0XxR]6DQGRYDO$):RRGFRFN ODQJXDJHSUR¿FLHQF\EDWWHU\²UHYLVHG²6SDQLVKIRUP ,WDVFD,/5LYHUVLGH3XEOLVKLQJ

Zill, N., Resnick, G., Kim, K., O’Donnell, K., Sorongon, A.,

+XEEHOO0F.H\5ZLWKRWKHUV+HDG6WDUW

)$&(6$ZKROHFKLOGSHUVSHFWLYHRQSURJUDP SHUIRUPDQFHU.S. Department of Health and Human 6HUYLFHV2I¿FHRI3ODQQLQJ5HVHDUFKDQG(YDOXDWLRQ Administration for Children and Families. Washington, '&86*RYHUQPHQW3ULQWLQJ2I¿FH