journal of literacy research 1981 mcgee 145 56

21
7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56 http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 1/21 Journal of Reading Behavior 1981, Volume XIII,  No. 2 EFFECTS OF THE CLOZE PROCEDURE ON GOOD AND POOR READERS' COMPREHENSION Lea M. McGee  Louisiana State University, Department of Curriculum & Instruction, Baton Rouge, L !"#"$  %stract . The effects of a cloze rocedure develoed from transfer feature theor! of rocessing in reading on immediate and dela!ed recall of good and oor readers "ere studied. #as! cloze assages "ere generated so all cloze deletions "ere successfull! sulied $! aro%imatel! &'( of &8 students used in a norming  rocess. T"ent! third)grade good and t"ent! fifth)grade oor readers read and recalled an eas! cloze and a normal, non)cloze version of t"o e%ositor! assages. Recalls "ere scored ac) cording to an anal!sis of discourse rocedure. Results indicated that fifth)grade oor readers remem$er more than third)grade good readers immediatel! after reading and after a one)"ee* dela!. +oreover, fifth)grade oor readers remem$er more from reading an eas! cloze assage than from reading a normal assage in immediate recall. Imlications for the use of this cloze rocedure as an instructional techniue are discussed. Imroving reading comrehension is an imortant goal, if not the most imortant goal of ever! reading rogram. Therefore, the urose of man! activities in learning to read is to increase comrehension a$ilities. -ne techniue used to enhance comrehension a$ilities, the cloze rocedure, consists of having students read a assage "ith "ords deleted in a regular attern of ever! fifth "ord Bormuth, 19//0. tudents sul! the deleted "ords "hile reading so that the comleted Downloaded from jlr.sagepub.com by guest on January 16, 2016

Upload: jenivee-baniel-farrales

Post on 24-Feb-2018

219 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 1/21

Journal of Reading Behavior 

1981, Volume XIII, No. 2

EFFECTS OF THE CLOZE PROCEDURE ON GOOD

AND POOR READERS' COMPREHENSION

Lea M. McGee

 Louisiana State University, Department of Curriculum & Instruction, Baton Rouge, L

!"#"$

 %stract . The effects of a cloze rocedure develoed from transfer feature theor!of rocessing in reading on immediate and dela!ed recall of good and oor readers"ere studied. #as! cloze assages "ere generated so all cloze deletions "eresuccessfull! sulied $! aro%imatel! &'( of &8 students used in a norming

 rocess. T"ent! third)grade good and t"ent! fifth)grade oor readers read andrecalled an eas! cloze and a normal, non)cloze version of t"o e%ositor! assages.Recalls "ere scored ac)

cording to an anal!sis of discourse rocedure. Results indicated that fifth)grade oorreaders remem$er more than third)grade good readers immediatel! after reading andafter a one)"ee* dela!. +oreover, fifth)grade oor readers remem$er more fromreading an eas! cloze assage than from reading a normal assage in immediate recall.Imlications for the use of this cloze rocedure as an instructional techniue arediscussed.

Imroving reading comrehension is an imortant goal, if not the mostimortant goal of ever! reading rogram. Therefore, the urose of man! activitiesin learning to read is to increase comrehension a$ilities. -ne techniue used to

enhance comrehension a$ilities, the cloze rocedure, consists of having studentsread a assage "ith "ords deleted in a regular attern of ever! fifth "ord Bormuth,19//0. tudents sul! the deleted "ords "hile reading so that the comleted

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 2: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 2/21

 assage again ma*es sense. 2loze tas*s, although initiall! develoed to measurereada$ilit! Ta!lor,

19'30, are considered to imrove a$ilit! to use conte%t clues effectivel! 4enned!5 6eener, 19&30 and to rovide motivation for reading Bloomer, 19//7 eitzman5 Bloomer, 19/&0.

lthough the cloze rocedure is "idel! used in classrooms, there is conflictingevidence from research on its usefulness as a techniue for enhancingcomrehension. 6hile several studies have demonstrated that cloze tas*s roducesuerior comrehension Bloomer, 19/:7 Bloomer, 19//7 4enned! 5 6eener,19&30, other studies have found that cloze tas*s roduce no $etter comrehensionthan other 

reading tas*s chne!er, 19/'7 4eitzman 5 Bloomer, 19/&0.2onflicting results on the effectiveness of using cloze tas*s to increase

comrehension ma! $e due to common "ea*ness found in cloze researchJongsma,

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 3: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 3/21

146  Journal of Reading  Behavi!

19&10. n imlicit assumtion in studies of cloze effectiveness is that cloze tas*s"ill roduce changes in comrehension strategies or rocesses from those "hichoccur in normal reading. ;urther, these changes "ill roduce increasedcomrehension. o"ever, the studies revie"ed failed to ma*e e%licit "hichcomrehension strategies or rocesses that cloze tas*s are e%ected to change andho" cloze tas*s are to roduce those changes. nother imlicit assumtion in mostcloze research is that cloze tas*s roduce changes in comrehension over time. ;or e%amle, researchers have studied the effects on comrehension "hen cloze tas*s

"ere used in a training eriod of a fe" to several "ee*s chne!er, 19/'7 Bloomer,19/:7 eitzman 5 Bloomer, 19/&7 4enned! 56eener, 19&30. o"ever, it could $ethat cloze tas*s induce immediate changes in comrehension rocesses at the timeof reading rather than roduce more general changes in comrehension strategiesover time. -nl! one stud! Bloomer, 19//0 e%amined the effects on comrehensionafter a single e%osure to a cloze tas*. o"ever, the immediate effects oncomrehension "ere not measured7 comrehension "as onl! assessed ten "ee*safter comletion of the cloze e%ercise.

If cloze rocedures are to $e effectivel! used in classrooms as a techniue for 

increasing comrehension a$ilities, then more s!stematic e%amination of theeffects of cloze tas*s, $oth immediate and long term, on articular comrehension rocesses are needed than have $een rovided $! revious research. Therefore, the urose of the resent stud! "as to investigate some imlicit assumtions and tocorrect "ea*nesses found in revious studies. ;irst, the effects of a single e%osureto cloze e%ercises on comrehension of the assage immediatel! after reading "eree%amined. In addition, a one)"ee* dela!ed measure of comrehension "as alsota*en to assess an! lasting effects of rocessing differences. econd, assageretellings "ere used to measure comrehension. <revious research Bloomer, 19//0

emlo!ed uestions as a measure of comrehension "hich ma! not rovide enoughdetailed information to assess changes in comrehension rocesses. Results of recent research

+e!er, 19&'7 +e!er, 19&&7 Thornd!*e, 19&&7 +arshall 5 =loc*, 19&&)&87Tierne!, Bridge, 5 2era, 19&8)&90 suggest that anal!zing assage retellings for idea units recalled ma! $e a more sensitive measure of comrehension. ;inall!,comrehension rocesses "hich "ere redicted to $e affected $! cloze tas*s "eresecified using transfer feature theor! of rocessing in reading ;inn, 19&&)&80.

;inn 19&&)&80 conceived of transfer feature theor! of rocessing in reading to

e%lain from the standoint of linguistic theor! ho" correct "ords are sulied incloze deletions. o"ever, this theor! can also $e used to e%lain "hat rocessesare induced $! reading a cloze assage that ma! not $e e%erienced $! reading a

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 4: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 4/21

normal, non)cloze assage. ;inn suggested that the a$ilit! to sul! a "ordcorrectl! in a cloze deletion is a function of le%ical mar*ers and transfer features.>e%ical mar*ers are s!ntactic and semantic comonents of "ords "hich e%ressimlicit relationshis a "ord ma! have "ithin a sentence 4atz 5 ;odor, 19/?7;illmore, 19/80. ;or e%amle, @$achelor@ ma! have the le%ical mar*ers Anoun,Asingular, Aanimate, and A agentive, among others, in the follo"ing sentence

10 The $achelor "ashed his ne" car.

ome of the le%ical mar*ers of a "ord that has $een deleted in a sentence ma! $e sulied imlicitl! $! the conte%t of the sentence. If @$achelor@ "ere deletedfrom sentence 10, a reader imlicitl! "ould $e a"are that the deleted "ord "ould

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 5: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 5/21

 Effects of Cloze Procedure on Comprehension  14"

have the le%ical mar*ers Aagentive and A animate. These le%ical mar*ers "hich aresulied $! the conte%t of a sentence are transfer features 6einreich, 19//0.

ccording to transfer feature theor! in order to sul! correctl! a "ord thathas $een deleted in a cloze assage, readers rocess transfer features of the "ord asu$set of the "ordCs le%ical mar*ers0 through the conte%t of the sentence and inferthe "ordCs remaining le%ical mar*ers not included in the transfer feature su$set.These rocesses then roduce "ord recognition and allo" the "ord to $e suliedin the deletion.

;inn rovided suort for transfer feature and le%ical mar*er rocessing $!using transfer feature theor! to redict ho" eas! or difficult "ords are to sul!correctl! in cloze deletions. ccording to transfer feature theor!, eas! clozedeletions deletions "hich a high ercentage of eole fill "ith the correct "ord0"ould $e filled "ith "ords "hich have a great num$er of transfer features. Dsingthe "ordCs man! transfer features "ould ena$le readers to infer easil! the fe"remaining le%ical mar*ers, and the "ord "ould $e roduced easil!. Dsing astandard freuenc! inde%, te%t freuenc!, and assage difficult! as measures of amounts of transfer features and le%ical mar*ers, ;inn redicted "hich "ords in

BormuthCs 19//0 assages "ould $e easier for readers to sul! correctl!. ;innCsreanal!sis of BormuthCs data revealed that transfer feature theor! does roduce anaccurate rediction of eas! and difficult cloze deletions and rovides suort for the theor! that transfer features and le%ical mar*ers are rocessed in reading.

 Eotions a$out transfer features and le%ical mar*ers ma! rovide ane%lanation of "hat comrehension rocesses are affected differentl! $! readingcloze assages than $! reading normal assages. If a assage "ere roduced "ithonl! eas! cloze deletions, readers "ould $e li*el! to sul! most of the deleted"ords. The! "ould rocess transfer features and infer le%ical mar*ers throughoutthe assage in order to sul! the deleted "ords and ma*e the assage once againsensi$le. >oc*hart, 2rai*, and Jaco$! 19&/0 ostulated that highl! redicta$le"ords, such as those easil! su)

 lied in cloze deletions, are minimall! rocessed or attended to in normal reading.6hile the le%ical mar*ers of a "ord must $e rocessed if that "ord is to $esulied in a cloze deletion, >oc*hart, 2rai*, and Jaco$!Cs notions suggest thatthese le%ical mar*ers ma! onl! $e minimall! rocessed in normal reading.Therefore, a more thorough rocessing of inferred le%ical mar*ers ma! occur "henreading a cloze assage than "hen reading a normal assage. more thorough rocessing of le%ical mar*ers, e%ressing relationshis among "ords in the

 assage and sentences, ma! roduce more ela$orated and integrated sentencememories. The more ela$orated and integrated sentence memories ma! facilitatecomrehension of eas! cloze assages over normal assages. The resent stud!

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 6: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 6/21

investigated "hether comrehension of eas! cloze assages "as greater thancomrehension of normal assages.

-f interest to educators is not onl! investigating and facilitatingcomrehension of normal readers, $ut also investigating rocedures "hich imrovecomrehension for readers "ho do not succeed as "ell as normal readers. 2lozeresearch has not differentiated "hether cloze tas*s should enhance reading for oor as "ell as good readers. Dsing transfer feature theor! to roduce cloze assagesreuiring rocessing of transfer features and le%ical mar*ers suggests that cloze rocedures ma! $enefit

 oor readers more than good readers. <revious research indicates that oor readers rocess in a "ord)$!)"ord manner 2romer, 19&F7 -a*en, 6einer, 5 2romer,

19&17

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 7: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 7/21

1?8 Journal of  Reading Behavior 

=olin*off, 19&/7 4endell 5 ood, 19&9G. +ore thorough rocessing of le%icalmar*ers redicted to occur "hen reading cloze assages indicates that readers of cloze assages "ould $e more li*el! to rocess relationshis among "ords andsentences. Therefore, oor readers ma! rocess in a more integrated manner "hilereading an eas! cloze assage than research suggests the! rocess "hile reading anormal assage. u$seuentl!, oor readers "ould $e e%ected to $enefit morefrom reading eas! cloze assages than good readers. The resent stud! investigated"hether there is an interaction $et"een reading a$ilit! and assage condition.

research aradigm for e%amining differences $et"een good and oor readers"hich is eseciall! useful in the resent stud! has $een utilized $! =uthrie 19&30and Ta!lor 19&80. Hifferences among students on the same reading level $ut notthe same age "ere investigated rather than differences among eers "ho "ere noton the same reading level. This research aradigm is aroriate for.the resentstud! since the ease "ith "hich cloze deletions ma! $e filled correctl! is a functionof reading a$ilit! rather than age ;inn, 19&&)&80. =ood and oor readers of similar reading a$ilities should $e a$le to sul! cloze deletions similarl!. Therefore, goodand oor readers should rocess transfer features and le%ical mar*ers similarl!.

+#T-H

u$ects

T"o grous of students "ere used in this stud!. total of &8 rural schoolchildren "ere selected to form the first grou. This grou included ?F third)grade

students "ho "ere mem$ers of reading grous la$eled $! their teachers as a$ovegrade level and 38 fifth)grade students "ho "ere mem$ers of reading grousla$eled as $elo" grade level. These students reresented all the third)graders"ithin an elementar! school considered $! their teachers to $e good readers and allthe fifthgraders considered to $e oor readers. ;ifth)grade students udged $! their teachers to $e reading $elo" the third)grade level "ere not included in this grou.This grou "as used to determine ho" eas! each "ord in the e%erimental assages "as to su)

 l! correctl! in a cloze deletion.

total of ?F students "ere selected from t"o additional elementar! schools toform the second grou. T"ent! students "ere randoml! selected from allthirdgraders reading on a 3.F)?.' grade level third)grade good readers0. T"ent!

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 8: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 8/21

additional students "ere randoml! selected from all fifth)grade students reading ona 3.F)?.' grade level fifth)grade oor readers0. Reading grade levels "eredetermined $! the reading su$test of the cience Research ssociates chievementTest <rimar! II0 administered "ithin four months of the time of the stud!. Thisgrou "as used to o$tain retellings of the eas! cloze and normal assages.

Materials

T"o 1:')"ord e%ositor! assages "ere constructed $! the investigatornimal Teeth and eeds. The t"o assages "ere segmented into roositions"hich "ere lacedCin a hierarchical structure using an anal!sis of discourse

 rocedureTa!lor, 19&80 adated from +e!er 19&'0 and =rimes 19&'0. 6ithin each roosi)

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 9: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 9/21

 ffects of Clo'e <rocedure on 2omrehension 1?9

tion redicates and their arguments "ere considered to $e searate idea units. The assages "ere "ritten to $e as similar as ossi$le "ith a third)grade reada$ilit!level ;r!, 19&&, an identical structural hierarch!, and a similar num$er of ideaunits /1 and /:0.

To o$tain the eas! cloze version of each assage, the ercentage of students"ho correctl! filled each "ord in the assage "hen it "as resented as a clozedeletion "as calculated. ince a deletion attern of seven intervening "ords "asused in the resent stud!, eight cloze te%ts of each assage "ere generated so thatthe deletion atterns started from different "ord ositions. In this "a!, the eightcloze te%ts allo"ed each "ord in the assage to $e deleted once.

The first grou of &8 students "ere administered these cloze te%ts in grous of eight to ten. The students first read and sulied deleted "ords in a ractice cloze assage. The investigator stressed that students "ere to feel free to guess, s*i

deletions and return to them, and as* for hel selling "ords. These instructions"ere reeated throughout the testing rocedure. Ee%t, students "ere assigned acloze te%t $oo*let at random. #ach cloze te%t $oo*let contained one cloze versionof each assage. -ne half of the students received the cloze version of nimal(eet) first, and the other half received the cloze version of eeds first. 2loze te%ts"ith different starting ositions "ere randoml! assigned "ithin $oo*lets. Eine or ten students comleted each of the eight cloze te%t versions of each assage. The ercentage of students "ho correctl! filled each deletion "as calculated. -nl!e%act "ords, ronoun su$stitutions for referent nouns, or noun su$stitutions for 

referent ronouns "ere acceted in calculation of the ercentages of correctl! filleddeletions.

The eas! cloze version of each assage "as generated so that at least /F( of the students correctl! filled each "ord deleted in the assage. s much as "as ossi$le, a seven intervening "ord deletion attern "as maintained in each cloze assage. nimal Teeth had a mean deletion attern of /.93 intervening "ords "itha mean ercentage of students correctl! filling each deletion of &?.93(.  See*s hada mean deletion attern of &.:1 intervening "ords and a mean ercentage of students correctl! filling each deletion of &?.8&(. The t"o assages are resented

in ;igure 1 "ith the "ords deleted in the eas! cloze assage condition in $old facet!e.

<rocedures

tudents "ere tested individuall! $! the investigator in order to o$tain assageretellings. ;irst, each student read a short ractice cloze assage filling in thedeleted "ords. Ee%t, each student silentl! read a assage filling in the deletions if the assage "as in the eas! cloze condition0 and recalled it orall!. Then, the studentsilentl! read the second assage and recalled orall!. The investigator stressed thatstudents could s*i "ords in the cloze assage and return to them, should feel freeto guess, and could as* for hel selling "ords. 6hen the students stoed their oral recall, the! "ere as*ed once if there "as an!thing else the! could remem$er.-ne "ee* later the titles of each assage in the order of original resentation "ere resented and each student again orall! recalled. -ne half of the students read thecloze version of nimal Teeth "hile the other half read the cloze version of eeds.6ithin each of those grous, one half of the students read the cloze te%t first "hilethe other half read the normal te%t first.

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 10: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 10/21

1'F Journal of  Reading Behavior 

Seeds

 ll !inds of  seeds are carried to new "omes in different ways. Some  seeds arecarried to t"eir new "omes by t"e wind. #"e wind can blow dandelion seeds acrossthe grass to some ot"er yard. $f t"e wind blows ot"er seeds to a good spot wit" lotsof sun and water, new trees or new plants will  grow. Seeds can be carried to t"eir new homes by furry animals or people. Some seeds may "ang in animals% fur. &r seeds may stic! in t"eir paws. Some seeds may catc" in peoples% clot"es. 'irds cancarry seeds and nuts to new "omes. $f birds drop seeds or nuts t"ey li!e to eat onroc!s, then a new plant or new tree will not grow t"ere.

 nimal #eet"

Many kinds of animals use t"eir teet" in different ways. Some animals use t"eir frontteet" to cut t"ings. (ots of bea)ers cut down trees wit" t"eir strong front teet". *abbits use t"ese teet" to cut t"eir food like carrots. nimals use ot"er teet" to tear t"ings. +"en ananimal li!e a bear  or  a wolf wants to tear its food, it uses t"e teet" at the side of its mout"wit" t"e points. Dogs and cats also use t"ese teet" to tear t"ings. nimals use still ot"er  teeth to c"ew t"eir food. +"en cows c"ew the grass t"ey eat, t"ey use t"e teeth in t"e bac!of t"eir mout"s. orses also use t"ese teet" to c"ew t"e "ay t"at they eat.

;igure 1. #%erimental <assages.

coring <rocedures and nal!ses

Recalls "ere scored for the ercentage of Idea units retold using an anal!sis of

discourse rocedure Ta!lor, 19&80. T"o ersons trained in the rocedure scored arandom samle of 1F&o of the recalls and o$tained inter)rater relia$ilit! of .98.

T"o t!es of te%t)$ases "ere used in the scoring rocedures to calculate ercentages of idea units retold, and t"o sets of anal!ses "ere conducted on these ercentages. The second set of te%t)$ases andCanal!ses "ere used $ecause readersof cloze versions of a assage ma! not $e a$le to form te%t)$ases "ith all ideaunits of a assage "hile readers of normal versions of assages "ould $e a$le toform te%t$ases "ith all idea units. That is, some readers ma! not $e a$le to sul!all the "ords deleted in a cloze assage. 6ords that readers fail to sul! in cloze

deletions "ould not $e resent in the te%t. The idea units sulied $! these "ordscould not $e encoded and could not $e included in the assage te%t)$ase. Incontrast, these "ords "ould $e resent in a normal te%t, and these idea units could $e encoded and included in the te%t)$ase. Hifferences in recall $et"een eas! clozeand normal assages ma! $e due to differences in a$ilit! to sul! cloze deletionscorrectl! roducing differences in idea units encoded rather than due to differencesin rocessing le%ical mar*ers. ince the first scoring method using the full te%t) $ase did not control for a$ilit! to sul! cloze deletions and for the num$er of idea

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 11: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 11/21

units encoded in a te%t$ase, the second scoring method using a modified te%t $ase"as used "hich did control for these factors.

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 12: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 12/21

 ffects of Clo'e +roce*ure on Compre)ension 1'1

;irst, a studentCs full te%t)$ase ercentage of idea units retold "as calculated $!dividing the num$er of idea units retold $! the total num$er of idea units in the fullte%t)$ase /1 and /:0. The the studentCs modified te%t)$ase ercentage of idea unitsretold "as calculated $! dividing the num$er of idea units retold $! the num$er ofidea units in hisKher modified te%t)$ase. The modified te%t)$ase for each studentCsretelling of a cloze assage included onl! the idea units resent in the te%t of thecloze version of the assage '3 and '?0 and all idea units roduced $! the students"hen heKshe correctl! filled a cloze deletion. earate anal!ses and ost hoccomarisons "ere conducted "ith modified and full te%t)$ase ercentages. In ever!case the significance of results "as unchanged using the modified te%t)$ase scoringmethod from that o$tained using the full te%t)$ase scoring method. Therefore, onl!results of the full te%t)$ase scoring method are reorted.

further e%amination of studentsC a$ilit! to comlete cloze deletions correctl!"as conducted. ince the norming grou from "hich the eas! cloze deletions "eregenerated "as not the same as the grou from "hich the assage retellings "ereo$tained, the ercentage of deletions correctl! filled for the latter grou of students"as calculated. The urose of these calculations "as to determine if the deletions"ere indeed as eas! for the recall grou as for the norming grou and to determineif fifth)grade oor and third)grade good readers filled cloze deletions euall! "ell.

R#D>T

: reader good and oor0 % : recall condition immediate and dela!ed0 % :

assage condition eas! cloze and normal0 mi%ed anal!sis of variance "asconducted "ith recall condition and assage condition reeated factors. There "asa significant main effect for reader ;i, n/ L 11/.'F, M .FF'0 and for recallcondition ;i.ne L 1&:.:8, M .FF10. Ta$le 1 illustrates that fifth)grade oor 

readers retold ideas $etter than third)grade good readers in immediate and dela!edrecall. There "ere t"o significant interactions reader $! recall condition ;i, n / L'.F&, M .F30 and assage condition $! recall condition  -', ne L /.:F, M .FF30.s sho"n in Ta$le :, the e%ected interaction $et"een reading a$ilit! and assagecondition oc)

Page 13: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 13/21

TABLE #

Mea$% a$& Sa$&a!& Deviai$% ( G& a$& P!

Rea&e!%' P!)!i$ae Ree##i$*%

Reca## C$&ii$ Thi!& G!a&e Fi(h G!a&e

G& Rea&e!% P! Rea&e!%

I++e&iae Reca## , .1-4 ./""0

SD .1" .16

De#a2e& Reca## , .3/ .14/0

SD .116 .11/

Mea$ &i((e!e$ce %i*$i(ica$

5) .3Downloaded from  jlr.sagepub.com by guest on January 16, 2016

Page 14: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 14/21

1': Journal of Reading Behavior 

TABLE

Mea$% a$& Sa$&a!& Deviai$% ( G& a$& P# Rea&e!%

P!)!i$ae Ree##i$*% ( Ea%2 C#7e a$& N!+a# Pa%%a*e%

I++e&iae Reca## De#a2e& Reca##

Rea&e! Ea%2 C#7e N!+a# Ea%2 C#7e N!+a#

Thi!& G!a&e , .1-1 .11 .3/" .3-

G& Rea&e!% SD .14/ .1/3 .11- .13

Fi(h G!a&e , .1 ./30 .14 .11

P! Rea&e!% SD .13 .1- .1- .11"

5Mea$ &i((e!e$ce %i*$i(ica$ )

.3

curred in immediate recall. HuncanCs +ultile Range Tests demonstrated thatfifthgrade oor readers recalled cloze assages $etter than normal assages M .F'0 "hile third)grade oor readers did not recall eas! cloze and normal assagesdifferentl!. This interaction is grahicall! disla!ed in ;igure :. Ta$le : alsoillustrates that in dela!ed recall the e%ected interaction $et"een reading a$ilit! and assage condition did not occur. Eeither fifth)grade oor nor third)grade goodreaders recalled eas! cloze and normal assages differentl!.

#%amination of the ercentage of cloze deletions filled correctl! $! good and oor readers revealed that fifth)grade oor readers correctl! sulied more clozedeletions 8F.8 / than third grade good readers /8.: 0. t)test used to e%aminethe mean difference demonstrated significance t3s L '.?1, M .FF10.

HI2DI-E

The urose of this stud! "as to rovide information for effective use of thecloze rocedure in increasing comrehensio.n a$ilities. Recall of e%ositor! te%t

"as investigated in the resent stud! since remem$ering from te%t $ecomesincreasingl! more imortant as students rogress through the elementar! gradesand $e!ond. Therefore, develoing techniues, such as the cloze rocedure, "hich

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 15: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 15/21

imrove learning and remem$ering information from reading is of rimar!imortance to educators.

=ood and oor readersC a$ilit! to comrehend and recall eas! cloze and normal assages "as investigated. ;ifth)grade oor readers remem$er more informationoverall than third)grade good readers immediatel! after reading and after a one)"ee* dela!. s e%ected, fifth)grade oor readers recall more ideas immediatel!after reading an eas! cloze assage than immediatel! after reading a normal assage "hile third)grade good readers do not recall eas! cloze or normal assagesdifferentl!. o"ever, these differences are not sustained in dela!ed retellings. Indela!ed recall neither grou of readers recall eas! cloze assages $etter thannormal assages.

Hifferences in recall among fifth)grade oor and third)grade good readers arenot attri$uta$le to different amounts of information availa$le at encoding. lthough

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 16: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 16/21

Efectso ClozeProcedureon Comprehension 153

 Third Grade Good Readers

Fith Grade Poor Readers

2

•a

o

c

o

s

Easy Cloze Passage ormal Passage

Figure !" Good and Poor Readers# Proportionate Retellings o Easy Cloze and ormal Passages

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 17: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 17/21

1'? Journal of Reading Behavior 

fifth)grade oor readers did comlete more cloze deletions than third)grade goodreaders and su$seuentl! ma! have had more information availa$le at encoding,the modified te%t)$ase scoring method controlled for this difference in the anal!sis.Because there "ere no differences in modified and full te%t $ase anal!ses,differences $et"een fifth)gradersC recall and third)gradersC recall are not due todifferences in amount of ideas availa$le in the te%t at the time of reading.

-lder oor readersC suerior recall overall ma! $e due to several factors. Thesereaders ma! have more *no"ledge a$out the assage toics, more e%erience

learning and remem$ering from e%ositor! te%t than third)graders, and morefacilitation roducing oral recalls. In addition, since fifth)graders correctl! filledmore cloze deletions than third)graders, then the assages must have $een easier toread for fifth)graders. The fact that the assages "ere easier to read ma! mean thatmore caacit! "as availa$le to the fifth)graders for learning "hile reading.

;ifth)grade oor readers $enefited from reading eas! cloze assages ase%ected. <rocessing transfer features and le%ical mar*ers, "hich according totransfer feature theor! is reuired in order to fill cloze deletions, ma! haveimroved "ord$!)"ord reading suggested to $e detrimental to oor readerCs

normal reading. Nounger good readers ma! not have recalled eas! cloze assages $etter than normal assages for several reasons, the! ma! have less e%erienceusing cloze rocedures than fifth)graders. lso, since third)graders sulied fe"er cloze deletions than fifthgraders, third)graders ma! not have rocessed transfer features and le%ical mar*ers as "ell as fifth)graders. This result "as une%ected.Nounger readers in the resent stud! "ere chosen so that their reading a$ilities"ere similar to those of older 

readers. ince the a$ilit! to fill cloze deletions correctl! is $ased on reading a$ilit!,these t"o grous "ere e%ected to erform similarl! on cloze tas*s. o"ever,

reading level in the resent stud! "as assessed $! a test of general readinga$ilities. Nounger readers, "hile having similar general reading a$ilities as older readers, ma! have $een less familar "ith the secific toics and s!nta% found inthe e%erimental assages. These factors ma! have influenced their a$ilit! to fillcloze deletions and to rocess transfer features and le%ical mar*ers. ;urther research ensuring that !ounger good readers are euall! as successful comletingcloze tas*s as older oor readers and, su$seuentl!, euall! as successul rocessing transfer features and le%ical mar*ers is needed $efore it is ossi$le toconclude "hether good readers "ould $enefit from reading eas! cloze assages asmuch as oor readers.

lthough fifth)grade oor readers erform $etter recalling eas! cloze thannormal assages immediatel! after reading, this result "as not reeated after a

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 18: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 18/21

one"ee* dela!. This ma! $e due to oor overall erformance in dela!ed recall 1'&o0. Indeed, $oth fifth)grade oor and third)grade good readers even in immediate

measurement recall less than one third of the information in the assages. -ther studies Ta!lor, 19&87 Tierne!, Bridge, 5 2era, 19&8)&97 Hi%on, 19&90 have alsofound that students recall similarl! small amounts of information after reading andlistening. It

ma! $e that students are still struggling to read and understand material and havelittle remaining cognitive caacit! to learn and remem$er. In articular, studentsma! find e%ositor! te%t difficult to recall. The! ma! $e ine%erienced reading for the detailed information found in this t!e of te%t, less familiar "ith the toics resented in this te%t, and less familiar "ith the organizational structures found in

e%ositor! te%t.

Downloaded from jlr.sagepub.com by guest on January 16, 2016

Page 19: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 19/21

 ffects of Clo'e +roce*ure on Compre)ension  1''

limitation of the resent stud! is that onl! short e%ositor! te%t recall "as investigated.ince material to $e learned in school is usuall! more length! than 1:' "ords, furtherresearch e%amining recall of long e%ositor! te%t is indicated.

In conclusion, the results of the resent suggest that "hen cloze te%ts are carefull!constructed so that all deletions are easil! sulied "ith correct "ords, these te%ts ma! $eeffectivel! used to increase comrehension. This techniue ma! $e eseciall! $eneficial for  oor readers. ;uture research should $e directed at e%amining effects of long)term training"ith eas! cloze assages on the comrehension of $oth good and oor readers.

R#;#R#E2#

B>--+#R, R. The cloze rocedure as a remedial reading e%ercise. 0ournal of Developmental Rea*ing, 19/:, 1, 1&3)18:.

B>--+#R, R. The effects of non)overt reinforced cloze rocedure uon readingcomrehension. In =. chic* and +. +a! #ds.0   Ne -rontiers in College3*ult  Rea*ing. +il"au*ee, 6I Eational Reading 2onference, 19//.

B-R+DT, J. Reada$ilit! a ne" aroach. Rea*ing Researc) 4uarterly, 19//, 5, &9)13:.2R-+#R, 6. The difference model a ne" e%lanation for some reading difficulties. 0ournal 

of *ucational +syc)ology, 19&F, 65, ?&1)?83.

HIX-E, 2. Te%t t!e and childrenCs recall. In +. 4amil and . +oe #ds.0 Rea*ing Researc)

Stu*ies an* pplications, 2lemson, 2 Eational Reading 2onference, 19&9.

;I>>+-R#, 2. The case for case. In #. Bach and R. arms #ds.0 Universals in Linguistic()eory. Ee" Nor* olt, Rinehart and 6inston, 19/8, 1)88.

;IEE, <. 6ord freuenc!, information theor!, and cloze erformance a transfer feature theor!of rocessing in reading. Rea*ing Researc) 4uarterly, 19&&)&8, 13, 'F8)'3&.

;RN, #. ;r!Cs reada$ilit! grah clarifications, validit!, and e%tension to level 1&.  0ournal of Rea*ing, 19&&, 25, :?:):':.

=->IE4-;;, R. comarison of reading comrehension rocesses in good and oorreaders. Rea*ing Researc) 4uarterly, 19&')&/, 55, /:3)/'9.

=RI+#, J. ()e ()rea* of Discourse. The ague, Eetherlands +outon and 2oman!, 19&'.

=DTRI#, J. Reading comrehension and s!ntactic resonses in good and oor readers.  0ournal of *ucational +syc)ology, 19&3, 61, :93):99.

#ITO+E, ., 5 B>--+#R, R. The effect of non)overt reinforced cloze rocedure.   0ournal of Rea*ing, 19/&, 55, :13)::3.

J-E=+, #. ()e Clo'e +roce*ure as a (eac)ing (ec)ni7ue. Ee"ar*, H# InternationalReading ssociation, 19&1.

4TO, J., 5 ;-H-R, J. The structure of semantic theor!. In J. 4atz and J. ;odor #ds.0 ()eStructure of Language. #ngle"ood 2liffs, EJ <rentice all, 19/?, ?&9)'18.

Page 20: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 20/21

4#EH#>>, J., 5 --H, J. Investigating the relationshi $et"een comrehension and "ordrecognition oral reading anal!sis of children "ith comrehension or "ord recognition

disa$ilities. 0ournal of Rea*ing Be)avior, 19&9, 55, ?1)?8.4#EE#HN, H., 5 6##E#R, <. Visual and auditor! training "ith the cloze rocedure to

imrove reading and listening comrehension. Rea*ing Researc) 4uarterly, 19&3,  #,':?)'?1.

>-24RT, R., 2RI4, ;., 5 J2-BN, >. Heth of rocessing in recognition and recallsome asects of a general memor! s!stem. In J. Bro"n #d.0  Recognition an* Recall ,>ondon 6ile!, 19&/.

+R>>, E., 5 =>-24, +. 2omrehension of connected discourse a stud! into therelationshi $et"een thDownloaded from e structur e  jlr.sagepub.co  of  te%t m  by guest on January 16, 2016andinformation recalled. Rea*ing Researc)

4uarterly, 19&8)&9, 58, 1F)'/.

Page 21: Journal of Literacy Research 1981 Mcgee 145 56

7/25/2019 Journal of Literacy Research 1981 Mcgee 145 56

http://slidepdf.com/reader/full/journal-of-literacy-research-1981-mcgee-145-56 21/21