Phonological similarity effects in simple and complex word spans

Appendix H: Low context sentences used in Experiment II……….143 List of Tables Table 1: Average means and standard deviations for word lists used in Pilot study………38 Table 2: Means of Abs

TABLE OF CONTENTS

Acknowledgement……… 6

Abstract ……… 7

Chapter I: Introduction .8

Working memory theory……… …….8

Some phenomena found in working memory………13

Simple and complex span tasks ……….15

Phonological similarity effect ……… ……19

Study by Copeland and Radvansky……… ………… 22

Context Availability Model……… ………… 25

Motivations for the current study……… 32

General Predictions……… ……… 35

Chapter II: Pilot Study …….……… ……… 36

Method……….…….….37

Participants……….………….… 37

Design….……… ……….37

Material….……….37

Procedures……….……… 39

Scoring……….….41

Results……… ………43

Error Analysis……… 46

Discussion……… ……… 49

Chapter III: Experiment I……….53

Method……….53

Participants……… 53

Design ……….…… .53

Material….……….… 54

Procedure……….55

Apparatus and Scoring……… ……… 57

Results……… 57

Analysis across trials……… 61

Discussion……… 64

Chapter IV: Experiment II……… 68

Method……….69

Participants……… 69

Design……… 70

Material……….70

Procedure……… 74

Scoring……… 75

Results……… 75

Discussion……… 78

Chapter V: General Discussion……… 81

Summary of results for experiment………81

Error Patterns……….84

Scoring procedure……… 85

Reversal of the phonological similarity effect……… 87

Contextual value of sentences……… … 89

Relevance of the findings for reading span task……… 90

Working Memory……… 90

Limitations of the current study and future directions……… 91

Conclusion………94

References……… 95

Appendixes……… ……… ….101

Appendix A: Word lists used in Pilot study……… …101

Appendix B: List of words changed for Experiment I………….……… 112

Appendix C: Word list used for word span task in Experiment I 113

Appendix D: Sentences for reading span task in Experiment I 122

Appendix E: Sample questionnaire used for Experiment II ………… …… … 133

Appendix F: Final questionnaire used for Experiment II 136

Appendix G: High context sentences used in Experiment II……….138

Appendix H: Low context sentences used in Experiment II……….143

List of Tables

Table 1: Average means and standard deviations for word lists used in Pilot study………38

Table 2: Means of Absolute span scores of recall for similar and dissimilar conditions in pilot study………44

Table 3: Means of Total span scores of recall for similar and dissimilar conditions in pilot study

………45

Table 4: Error analysis Pilot study ……… 48

Table 5: Average means and standard deviations for word lists used in Experiment I… 55

Table 6: Means similar and dissimilar conditions in word and reading span tasks……….57

Table 7: Error analysis Experiment I……… ………60

Table 8: Average means and standard deviations for word lists used in Experiment II….70

Table 9: Serial recall proportion across list type and context conditions ……….75

Table 10: Error analysis Experiment II……….77

List of Figures

Figure 1: A simple representation of the working memory model by Baddeley and Hitch 10

Figure 2: Means similar and dissimilar conditions with Total span scores……… …44

Figure 3: Means similar and dissimilar conditions with Absolute span scores….……… 46

Figure 4: Means for word and reading span using Absolute span task……… … 58

Figure 5: Means for word and reading span using Total span task……… 58

Figure 6: Analysis across list lengths for word span………62

Figure 7: Analysis across list lengths for reading span………63

Figure 8: Mean proportion of recall of similar and dissimilar words……… 76

Acknowledgement

I would like to thank my supervisor, Dr Winston Goh for his immense and valuable help and

support His guidance has helped me achieve a deeper knowledge of the subject

Also, I would like to extend my sincere gratitude to David Copeland for sending me the samples of the materials used in his study

A few other people deserve to be thanked for their assistance in preparing this thesis My sincere thanks to the reviewers who provided good feedback which helped me shape this study Also I would like to mention my parents who gave me support in every way possible

Lastly, I would like to extend my sincere gratitude towards National University of Singapore for giving me an opportunity to put my thoughts and ideas to practice

Abstract

Memory span for a list of phonologically similar words is generally worse than memory span for a

list of phonologically dissimilar words, a finding that is called the phonological similarity effect

This finding has often been cited as evidence for the use of phonological coding in short-term

memory and working memory However, some studies have demonstrated a reversal of the

phonological similarity effect under certain conditions One such condition is the use of more

complex memory span tasks such as reading span It has been suggested that sentence contexts may provide additional retrieval cues that may overcome the detrimental effects of phonological

similarity The present study examined this hypothesis by manipulating the sentence contexts of the reading span materials A pilot study confirmed the phonological similarity decrement in a word span task, showing a higher recall for dissimilar sounding words than the similar sounding words This was followed by Experiment I which consisted of a word span and a reading span task The results showed that unlike the word span task, phonological similarity facilitation effect was

observed in the reading span task Further extension of the study led to Experiment II which tried to investigate the effect of context on the recall of the participants on a reading span task employing high and low context sentences No evidence showing phonological similarity facilitation was found;

in fact, the standard phonological similarity decrement in memory span performance was observed with high context sentences

Chapter I: Introduction

Working memory theory

The expression ‘working memory’ was initially used by Miller, Galanter and Pibram (1960)

in the book titled ‘Plans and the structure of behavior’ According to the authors, the study of

working memory and its concepts are recognized as a milestone in the early development of

cognitive psychology, and interestingly working memory was assigned a central role to this account

Working memory collects different information in the mind of the individual at one point of time and then relates it to other things It could be defined as a system which is responsible for the brief storage and management of the information essential for accomplishing complex cognitive actions like learning a task, reasoning out things, and comprehension of text, etc (adapted from MedicineNet.com) This system could thus be understood to be involved in the selection, initiation, and termination of information-processing functions such as encoding, storing, and retrieving data Such a system seems to be very helpful if one is striving to read sentences whose meaning can only

be deciphered after it has been fully read (Baddeley, 1999) Other tasks which require a similar system are the processes of performing mental mathematics, reasoning out things, and also times when one is faced with a task of problem solving (Baddeley, 1999)

The roots of working memory can be traced from the theories of short-term memory

(Andrade, 2001; Baddeley & Hitch, 2000) which primarily focused on the storage of information on

a temporary basis Discussion regarding the role of short term memory has more often concluded the

fact that short term memory functioned as working memory (Andrade, 2001) In contemporary

theories of working memory, the theory of Miller et al., (1960) is not widely mentioned Traces of

origin of the concepts of working memory could be found in the studies by Atkinson and Shiffrin (1968) Their study contained a thorough analysis of views related to the structure and the functions

of human memory The analysis of their study postulated three main components involved in the structure They were a sensory register, a short term store, and a long term store Atkinson and Shiffrin were of the view that the short term store could be considered to be ‘working memory’ which receives inputs from the sensory register and the long term store

Both the original accounts presented by Miller et al (1960) and that by Atkinson and Shiffrin (1971) demonstrated what Baddeley (1986) later defined as the ‘general concept of working

memory’ Baddeley and Hitch (1974) identified three main components of working memory as depicted in Figure I

CENTRAL EXECUTIVE

Figure1: A simple representation of the Working Memory Model by Baddeley and Hitch Adapted from Baddeley (1999)

The central executive component is the most important one and its functions included regulation

of information within working memory, and retrieving of information from other systems like long term memory and the processing and storage of information The cognitive tasks that appear to involve the central executive include mental arithmetic, recall of digits, logical

reasoning, semantic verification and recollection of events from long term memory

There have been a number of advances in understanding the central executive component of working memory (Baddeley & Logie, 1999) Much of the current work on the regulatory functions

of the central executive is guided by a model developed by Shallice (1982, see also Norman & Shallice, 1986) In their model action is controlled by two ways Well learned or automatic activities

are guided by schemas that are generated by environmental cues According to their study, the supervisory attention system is a limited capacity system and is used for a variety of purposes, including tasks involving planning or decision making, trouble shooting in situations in which tautomatic processes appear to be running into difficulty, novel situations, dangerous or technically difficult situations, situations where strong habitual responses or temptations are involved

ystems’ Each slave system is specialized for the temporary maintenance and the processing

of the material within a particular area One of the two slaves is the phonological loop which is responsible for maintaining verbally coded information The phonological loop comprises two components, the phonological store and the articulatory loop The phonological store includes tmaterial in a phonological code and these codes decay as time passes (Baddeley & Gathercole, 1993) A process of rehearsal using the articulatory loop helps refresh the decaying representatio

in the phonological store and also maintains memory items

C

temporary storage of verbal information, particularly novel phonological input (Baddeley, Gathercole, & Papagno, 1998) Yet another important function which it performs is that the

temporary storage of the material allows the listener to create long-term phonological represe

of that material (Baddeley et al., 1998)

encoded in the form of imagery Baddeley (1986) hypothesized that the visuo-spatial sketchpad is able to serve imagery as well as short term memory functions, arguing that there seems to be soevidence available to prove the occurrence of a visuo-spatial store (pg 143) This slave system specializes in the processing and storage of visual and spatial information It is also capable of processing and storing verbal information in the form of images (Baddeley & Gathercole, 1993Storage and processing of information are the two distinct aspects of working memory The cenexecutive is mainly responsible for the processing, control and integration of information whereasthe storage component is taken care of by the slave systems, although some simple processing mechanisms such as rehearsal in the articulatory loop are placed in the slave systems In terms of theoretical progress related to the functions and structure of the two slave systems, the phonololoop has made more progress than the visuo-spatial sketchpad It has been revealed that the

phonemic similarity effect has played a significant role in the investigation of the phonological loop

On the other hand, studies of visual similarity have not found equal importance in the studies

sketchpad as have the phonemic effect in the phonological loop (Baddeley & Hitch, 1994)

According to Baddeley and Hitch (1994), if it is understood that the sketchpad is a te

me

) tral

gical

of

mporary ace for the storage and manipulation of visual and spatial information, it gives a clue that it must sp

be accomplishing a number of tasks However, there exists scant research on this aspect of the spatial sketchpad It is expected that the sketchpad may be involved in activities like planning and executing spatial tasks, keeping track of changes in the visual perception of an individual and

visuo-comprehending certain types of visual and verbal information

Working memory has been one of the most explored areas of research (e g., Nairne, 1996) Variou

eld,

ome phenomena found in working memory

re and functions of the working memory, a few

S

In the process of exploring the structu

ena associated with storage and processing of information were identified (Andrade, 2

Articulatory suppression was one of the phenomena associated with working memory (Gathercole &

Baddeley, 1993; Mueller, Seymour, Meyer & Kieras, 2003; Russo & Grammatopoulou, 2003) According to Murray (1967) articulatory suppression refers to the procedure in which a participarequired to repeat a syllable aloud while performing a cognitive activity While doing this, his/her short-term memory for verbal material would be impaired According to the working memory theo(Baddeley & Hitch, 1994) the disturbing effect of articulatory suppression arises because it prevents the subjects from rehearsing the material Phonological recoding is disrupted and so the subvocal rehearsal is disallowed, with the result that the contents of the phonological store cannot be refreshand so will decay and be forgotten (Baddeley & Gathercole, 1993)

recall of lists containing five-syllable items (e.g., university, opportunity) Further word length effects in immediate memory were found both for memory lists presented visually and verbally (Baddeley & Gathercole, 1993)

Another observation was the irrelevant speech effect in which the recall of verbal material

or dialect (Baddeley, 1999) Some work related to this effect done is that of Colle and Welsh,

(1976), in which the presence of irrelevant foreign language noise reduced recall performance

phonologically different lists but not on phonologically similar lists Another study done by

Baddeley and Salame (1982) found that the use of spoken digits and other irrelevant words c

interfere with the digit span task Recent research done by Jones (1993) in this area has revealed ththe interfering material need not be speech but could also be any kind of sound that varies in its pitch

and Hull (1964) Baddeley (1966) and Conrad and Hull (1964) demonstrated that when a participant is presented similar and dissimilar sounding words, the similar words would be harder remember According to Baddeley (1990, p 72), the memory traces available for the similar

sounding material seem to be harder to discriminate at the time of recall The phenomenon ho

when a participant is faced with a task of memorizing a list of words (e.g., cat, mat and rat), letters (e g., C, D, B, and V) or even sentences containing similar sounding words

As the present study will focus on the phonological similarity effect, the other phenomena ssocia

imple and complex span tasks

e clearly categorized into two main types, simple span tasks and

S

Memory span tasks can b

x span tasks The simple span task (Baques & Saiz, 1999) is made use of when the research

simply intends to measure the memory span or storage capacity of the individual (e g., Fallon, Groves & Tehan 1999; Tehan, Hendry & Kocinski, 2001) Keeping in mind the motive of the taparticipants are given a simple cognitive task to perform and then their memory capacity was

measured with the help of their responses

ed to memorize a sequence of digits and then recall them in the same order as they were presented (serial recall) or in any order (free recall) Testing of the digits (Groth & Baker, 2003)generally began with a sequence of two digits for a variable number of trials after which the numb

of digits to be remembered is increased by one This general procedure has also been used with other material such as word lists (Postman, Turnage & Silverstein, 1964) and letters (e g., MacAndrew, Klatzky, Fiez, McClelland & Becker, 2002)

D

study th be

)

memory in hich they either have to remember letters, words or digits whereas in a complex span task like reading

This category of task was termed the complex span task The complex span tasks involve two tasks

which are to be performed simultaneously One task could be either reading a sentence, solving a mathematical operation or reading a comprehension passage whereas the second task required

participants to memorize a list of words, letters, etc (Copeland & Radvansky, 2001)

In a simple span task participants are only tested on the storage component of

w

span the participants were asked to read sentences and are simultaneously required to

remember a list of letters, words or digits Such a task would involve both the processing and stocomponents of memory, in which the participants would be involved in processing the sentenc

(processing element of working memory) and at the same time they would be required to memorize the list of words or digits (storage component of working memory) and recall them when requiredAlso the difficulty of the complex span tasks would be obvious from the fact that the memory traces

of the last word of the sentences suffer from resources channeled to perform the simultaneous task (reading the sentence) In other words, the participants performing the complex span task would not have all the resources at their disposal to perform any one task (either reading sentences or

memorizing words) Instead, the resources available to perform these two tasks would be divided

On the other hand, in a simple span task, there would be no such division of resources to pe

singe task, leading to channeling of all the available resources towards only one task (i.e.,

memorizing the words) It would thus imply that a more difficult simultaneous task like solving mathematical questions would results in lower recall scores because it would involve a lon

duration of processing of the material resulting in a greater decay (Towse, Hitch & Hutton, 1998)perhaps interference

There are a few

or

types of complex span tasks found in the studies done in this area One of the

most frequently used tests was the reading span task This test was devised by Daneman and

ter (1980) and according to them the aim of the complex span task was to tax both the

processing and storage functions of working memory rather than simply observe the working

storage functions as traditional digit span or word span tests would do Participants were given

sentences to read and simultaneously memorize certain list of words provided to them Their taskthen was to recall the last words of each of the sentences read (see also, Siegel, 1994; Daneman Tardiff, 1987) For example, if the participant was shown three sentences like:

Non performers who are considered useless are fired by the boss

For all humans living in the world shelter is a basic need

Participants were asked to read each sentence aloud Later when inst

Pointe & Engle, 1990; Linderholm & Broek, 2002) This type of test has also been used for other languages like Japanese (Kondo, Morishita, & Osaka, 1999)

Operation span is another test devised by Turner and Engle (1989) and La Pointe and Engle

990) They replaced the sentences in the reading span task with the verification of mathematical equatio

they

re

ict reading comprehension abilities of the individual and also that the simple and the complex word span tasks do not exactly reflect

incorrect When they verified the equation, they were shown words for example ‘men’ and then were shown another similar equation accompanied by another word When instructed, they we

asked to recall men and other words that followed it, in the same order

It was observed that simple span task (word span) could not pred

the same structures, processes or strategies of the memory of individuals (LaPointe & Engle, 1990) In the simple span tasks, the participants were only required to retain items (words, letters odigits) in short term memory and recall them whenever they were indicated to do so but in the case

of complex span tasks, the individuals had to perform complex cognitive processing simultaneously while memorizing some items shown to them Making use of a simple span task in any study is a good way to measure the storage component of the working memory of the participants However, in

a simple span task, the participants are only required to memorize either unrelated word lists or lis

of letters and digits; there was no requirement of any kind of comprehension in the way that reading

text passages or sentences would entail On the other hand in the complex span tasks, in which the participants are provided with sentences to read and mathematical equations to solve, the processing component required to parse these stimuli is engaged This means that complex span scores include

a component of working memory that would be engaged in reading and comprehension processes It

is therefore not surprising that reading span measures would also predict reading comprehension abilities

Th

d

e memory span procedure has been the most frequently used technique to estimate the emory capacity of any individual The development of complex span tasks to tap processes that may oth

The phonological similarity effect occurs even in the absence of any auditory input and the

ffect is sometimes also referred to as phonological similarity decrem

oop gives two explanations for the phenomenon One of the possible asons behind this effect could be that while rehearsing and recalling, the phonological codes of the

m

erwise not be taxed using simple spans suggests that previously established patterns of results using simple span tasks may not necessarily be replicated with complex span measures To explore any of the characteristics of working memory and its functions, the easiest and the mostconvenient way would be to assess the memory of the individuals with the use of both simple and complex span tasks (La Pointe & Engle, 1990)

Phonological similarity effect

presence of only visual inputs This e

ent (Copeland & Radvansky, 2001), signifying the general trend that similar sounding wordswill lead to poorer recall

The phonological l

re

already read material decay more quickly when they are similar (Posner & Konick, 1966) owingthe fact that phonological confusions arise from the similar sounding words This confusion leads to

an early decay of phonological codes and leads to worse recall compared to the dissimilar sounding words The second possible cause could be that the codes might decay at the same rate regardless of their similarity, but while recalling them, the partly degraded codes of phonologically similar items are very difficult to reconstruct and redintegrate (Nairne, 1990) Either of these possibilities would have possibly led to the lower accuracy of the recall of phonologically similar material in a serial recall task

As m

to

entioned earlier the phonological similarity effect has been a well explored

henomenon A number of studies have been conducted in the recent years to explore all the

possibl

sition same,

made use of both words and non words for nalyzing the existence of phonological similarity effect This effect was studied in the tasks of serial recall, including a serial recognition and item identification task The two serial recall experiments

p

e aspects and causes of this effect A study by Fallon et al., (1999) investigates the

phonological similarity effect in a serial recall task According to their study, “phonological

similarity is observed to affect the serial recall of material detrimentally when correct-in po

scoring is used” (Fallon et al., 1999, pg 301) They conducted two experiments to support the

in which they tested the participants using similar and dissimilar lists under immediate recall,

articulatory suppression and filled delay conditions The results revealed that similar word lists exhibited better item recall but as far as accuracy of the positions of the items was concerned,

dissimilar lists seemed to be better than the similar ones

In their study, Lian, Karlson and Winsvold (2001)

a

ilar lists

d Kelley (1999), it was shown that the ffect of phonological similarity decrement got reversed i.e., phonologically similar words recalled better t

ns is affected by lexicality of

e list items, type of phonological similarity and scoring methods.’ (Lian, Karlsen & Eriksen, 2004) The stu

the ial tudy

trated a significant interaction effect between lexicality (words vs nonwords) and

phonological similarity (phonological similar vs distinct lists) revealing that the sensitivity to

phonological similarity is higher for words than nonwords In addition to these results, it was also discovered that in the serial recall tasks, the differences in the scores of similar and dissim

appeared to be greater for the words than for the nonwords

Some recent studies have found that phonological similarity does not always lead to a

decrement in memory performance In the study by Nairne an

e

han dissimilar words, in the conditions when participants were given short periods of

distraction activities It was also observed that there was a better retention for the order of the items

in case of similar instead of dissimilar lists (Nairne & Kelley, 1999)

Another study which emphasizes the existence of the reversal of the phenomenon shows that

‘phonological similarity decrement in immediate serial recall conditio

th

dy reveals that the phonological similarity effect appeared in the serial recall score when phonologically dissimilar words were compared to words which share the middle vowel and the consonant in the end of the items (e.g., gub, dub, rub, etc) in the lists Whereas in the second

experiment of the same study it was found that the effect of phonological similarity was absent inprocess of serial recall when the dissimilar words were compared to the words that shared the initand final consonants (e g., gap, gop, gup) The materials used in the first experiment of this s

were similar to the similar words used in the current study, on the other hand the words used in the

second study were though not same as dissimilar words used in the current study but were not fully similar sounding (gAp, gOp, etc.), giving us an insight into the possibility that the phonological similarity effect is more likely to occur when the items share the rhyme component

The study by Copeland and Radvansky (2001) also found similar results when they obserthe reversal of the phonological similarity effect in the reading span task It has been

ved suggested that sentence contexts may provide additional retrieval cues that may overcome the detrimental effects of phonol

tudy compared the performance of participants on phonologically similar and

issimilar words, using word span and reading span tests In the first experiment, the participants

word span and reading span task with the use of similar

were tested for their memory capacity on a

similar word lists For the word span task, the participants were shown three sets of th

four, five, six and seven words At the end of each set the participants were required to recall the words aloud which were then typed by the experimenter The procedure was similar for the reading span test also, the only difference being that the participants were presented sets of sentences whicthey had to read aloud and then recall the last words of each of the sentence

The result for the first experiment indicated that in the word span tasks the recall of the participants exhibited the usual phonological similarity effect Participants sc

d

reading span task displayed an exactly opposite pattern The results demonstrated that the

participants could recall the similar words better than the dissimilar words presented to them

It was understood that the obtained results could have been found because a differe

of individuals were tested on the two different span tasks Perhaps, the results were due to indiv

nt group idual differences and different memorizing and recalling capacities of individuals In order to overcome this lim

in the reading span test contributed to the phonological similarity ading span tasks In this experiment, the researchers decided to test the particip

in the reading span tasks

The third experiment by Copeland and Radvansky (2001) aimed at reinforcing the idea that the content of the sentences

facilitation in the re

ants on the operation span task because operation span task did not contain any context related matter The participants were thus subjected to mathematical problems and were also

required to memorize the words shown to them Results revealed that the participants scored

in the dissimilar conditions than in the similar conditions, showing the traditional phonological similarity effect

Only the results for the reading span tests found a reversal of the phonological similarityeffect Copeland a

span w

The concert was held by a musician of international fame

The game was going on well when the storm came

The person extracted money from people by using a fake name

the retrieval cue could be coupled with the knowledge that the words in the set rhymed with each other However, in the dissimilar condition, for example,

While learning to swim the child drowned in the pool which was deep

The old man did not dispose his cigarette in the bin and dropped the butt

A person who does not like coffee can choose from the many varieties of tea

vansky, 2001; pg

sk even though it was a complex span task like the reading span In the case of operation span sks, the content of the mathematical operation was in no way related to the words presented to the particip

as the only task which included a retrieval cue that was semantically related to the target word In the similar condition, for example,

Radvansky, 2001; pg 776) The results obtained from the study clearly indicate a need for future

exploration into the reasons for such a pattern of results The word spans followed the already

existing trend of phonological similarity effect Similar pattern was observed in the operati

task too However it is not clear as to why the results for the reading span task were in a totally opposite direction

Copeland and Radvansky (2001) suggest that there is a possibility that the above mentiopattern of results ma

on span

ned

y be due to the cues emerging from the content of the sentences used in the reading span task In other words, the content of the sentences used in the reading span tasks provide some c

Context Availability Model

The stu

a

of complex span tasks like re

gnificance for studies related to the reading span tasks, studies based on phonological

similarity effect and its reversal and also the research based on the importance of context and its

availability This claim can be aptly supported by the context availability model, which provides a

n depends on the association of information and its contextual value The comprehension of the incoming

general model of comprehension” (Wattenmaker & Shoben, 1987, pg141)

The underlying view of the model is that the process of comprehensio

m l/information could possibly be derived from the material presented to the participant orthe participant’s knowledge of the material “Comprehension is successful to the degree t

comprehender is able to supply cognitive contributions necessary to understand the relationships among the concepts in the message and to form elaborate memory trace” (Wattenmaker & Shobe

1987, pg 141)

For example, in a condition when participants were provided with abstract context, they were shownsentences like:

There is

After a few minutes, however, they realize its difficulty and give up in frustration

At first impression, the challenge always seems to be

other hand, in a concrete context condition participants were provid

There is a tree that looks easy to climb, but the problem is that the branches are far ap

After a few minutes, however, they realize the problem and jump to the ground

At the first impression, the challenge always seems a trivial one

ts could not form very laborate memory traces obtained from the sentences compared to the concrete context sentences,

or

he

ben (1983), mentions “the context availability model as iscussed by Kieras (1978) was developed for the purpose of explaining the effects of the role of context

e

efore, it &

It can be seen that in the abstract context condition, the participan

problem being referred to in the sentences

A study by Schwanenflugel and Sho

d

s in sentences.” (pg 83) Also according to Schwanenflugel and Shoben (1983), “this model states that the comprehension processes in the language are aided in a very important way by the additional contextual information to the materials that are to be understood” (pg 83) It can thus be put that in the reading span tests, the contextual information is provided to the individuals from thenvironment of the stimulus As per the model, comprehension of the meaning of the sentences occurs when the person is able to make these kinds of cognitive contributions (Bransford &

McCarrell, 1974) On the other hand when individuals are not able to make the suitable cognitivecontributions, the sentences presented to the participants become meaningless for them Ther

becomes difficult for the participants to recollect the information already provided to them (MedinSchaffer, 1978)

The context availability theory mainly explains the effects in terms of the availability of context during comprehension for abstract and concrete materials The model of context availability lso mentions that individuals have difficulty determining appropriate contextual information for abstrac

predictions of this theory is that the general advantage of concrete material over abstract material occurs because concrete material readily provides more contextual information making it easier to identify and comprehend Though the model primarily mentions the existence of such a pheno

in the case of abstract and concrete material, a possibility that the same model could provide someinformation about the complex span tasks as well cannot be ruled out

Shwanenflugel and Shoben (1985) in their study indicate that if the material provided to theparticipants are inherently rich in their contextual information, the mate

d for the same purpose Similarly, in the complex span task and especially the reading span, there is a high chance that the sentences involving some variations in the contextual value (high context and low context sentences) would lead to some interesting results The high contextua

material will be comprehended better compared to the ones with low contextual value When

referred to as high context material, it is meant that the information provided to the person was hi

in its content value, easily comprehended whereas low context material would involve information

which is loosely put together and presented, having less meaning thus making the comprehension of the information very difficult

An effort to find out the significance of availability of context in the reading span task makes this study unique because it would be one of the earliest in the area which aims to explore, as there have be

e

o be

the role of the content of the sentences in the various aspects of the working memory capacity of individuals Under this category, some studies have focused

information

t

en several studies in the past (Wattenmaker & Shoben, 1987; Schwanenflugel & Shoben, 1985) which have emphasized the role played by context However, none of these studies have explicitly targeted the complex memory spans (especially reading span test) On the other hand, whether the contextual information provided with the high and low context sentences would havany interaction with the type of material (similar sounding words & dissimilar sounding words) tcomprehended would be exciting to know

Studies have previously investigated

on the context of sentences involving abstract and concrete material A study by

Wattenmaker and Shoben (1987) looks into the effects of the same on the recalling capacity of individuals in which experiments were performed to examine the influence of contextual

on the recall of abstract and concrete sentences They categorized the availability of context intothree, namely, random context, moderately coherent context and an only abstract or only concrete context conditions The results indicated that nearly in all the conditions, concrete sentences werebetter recalled than the abstract sentences In the study, the deficiency of the abstract words and theadvantage of the concrete words were mitigated by the context of the sentences used In the presenstudy, context availability will be manipulated with, low context sentences and high context

sentences For explaining the context availability model for the current study, it could be expected that the results obtained from this study would be similar to the study by Wattenmaker and Sh

(1987), when the abstract words condition is replaced by low context condition and the concrete words condition is replaced by high context sentence condition In other words, in both the studies (by Wattenmaker & Shoben, 1987; current study), context plays a major role in the comprehensio

of the sentences provided to the participants, whether it be the abstract and concrete or high and lowcontext Using context availability as a factor allows an investigation into whether the memory span

of the participants will be affected by the contextual value of the material provided to them

Another study by Marschark (1985) investigated the role of imagery in the sentences a

oben

n

nd the ading passages on recall In the 3 experiments performed in their study, recalling capacity of the particip

comprehensible (cf Schwanenflugel & Shoben, 1983) Concrete words and materials are better remembered than abstract materials because of the dual availability of semantic and percept

information is memory” (Marschark, 1985, pg 744)

Significant research done in this area examine

c

Similarly in a low constraint sentence:

e his birthday so his parents took him to the bar

ents which tried to examine the effect of sentence constraint in the form of high and low constraint and validity of cues on the processing of the expected and the unexpected sentence

completion The high constraint sentences were the ones, which had a high contextual constraint o

in other words the degree to which the participants will produce a word as a possible completiothe sentence On the other hand low constraint conditions comprised of lists of sentences with low contextual constraint or the degree to which participants will not choose a word for the most possiblecompletion of the sentence Participants received highly constraining and low constraining sentencefor the task This method used in the study by Schwanenflugel and Shoben (1985) is very similar to the method used to select the sentences in the last experiment of this study, discussed in chapter 4, where the participants were required to choose the most predictable word for the high and low context sentences for forming the list of sentences in experiment II The two conditions representedsituations in which the performance of the participants was observed using sentences with high low contextual values The contextual value of the sentences was manipulated to observe the

variations in the participants’ performance due to the high and low constraint in the context of sentences For example, high constraint sentences appeared to be something like:

The old person drank too much wine while sitting in the bar

The word “bar” would be highly expected within the sentence context

The child wanted to celebrat

The word “bar” is unexpected within this sentence context

and Shoben (1985) revealed that the sentences aided lexical decisions only for the expected completions, on the other hand, lo

, of

All the above mentioned studies though targeted a number of aspects related to the memory

em ventured into the complex memory span tasks and the areas related

vided ssibility

Results from the first experiment of Schwanenflugel

high constraint

w constraint contexts showed a lower context effect The current study employs a similar idea in which the similar and dissimilar sentences are further categorized into two categories eachhigh context and low context sentences as the material provided to the participants for the reading span task Similar technique has also been employed in one of the questionnaires (see Appendix F)

used for obtaining high and low context sentences for experiment II

Motivations for the current study

of individuals, however, none of th

to it Therefore, a need to investigate the role played by sentences’ context would be helpful

to answer the questions posed by Copeland and Radvansky (2001) in their study where they

established the effect of phonological similarity decrement in the word span task and when

extending the same to a reading span task it was observed that the effect of phonological similarity decrement had been converted to phonological similarity facilitation When the results were

investigated further on an operation span task instead of reading span it was noticed that the

phonological similarity effect appeared again, suggesting that the context of the sentences pro

some cues which aided the memory and recall of the participants The study discussed the po

of these contextual clues being made available to the participants as a future direction

Copeland and Radvansky (2001) did not specifically mention the mechanism by which the sentences’ context could reverse the phonological similarity effect In word span tasks, the

knowledge that all the words within a list rhyme may presumably provide a cue to limit recall to words t

One possibility is that the different sentence frames serve as distinctive cues that allow the last words to be better distinguished from each other or more salient The additional distinctive cue

s In other words, in reading sentences, people may encode the sentences and use them as retrieval cues, and coupled the knowledge of the rhyming organization of the to be remembered words, lead to better recall

The present study has two specific aims The first is to determine if the phonological

similarity effect reversal in th

chwanenflugel & Shoben, 1985)) have shown that people recall words better when the

embedded in high constraint or high context sentences, compared to words embedded in low

constraint or low context sentences (the words memorized could be found anywhere in the sentence

in the form of a fill in the blank) A more direct test of whether context is truly implicated in the

reversing the phonological similarity effect would be to manipulate the contextual value of thesentences in the reading span task

If the sentence context provides an effective retrieval cue, then sentences such as

ames reverses the detrimental effects of phonological similarity, then one ould expect the reversal to be greater when such cues are expected to be maximally effective in the

n the recall of lists of words provided to the participants The lists used for

e study would comprise of phonologically similar (mat, bat, cat) and dissimilar words (men, hit,

be very effective as a cue for retrieving the last word ‘bar’ because the word is high

expected within th

The child wanted to celebrate his birthday so his parents took him to the bar

Would not be a very effective cue for the last word ‘bar’ The word is unexpected and can be eareplaced by anot

Importantly, if Copeland and Radvansky’s (2001) hypothesis that contextual retrieval cues provided by the sentence fr

w

high context sentences

Thus, the main aim of the current study was to observe the effects of the difference in the context of the sentences o

th

he participants would be tested on the word span and reading span tasks The word span tawould comprise two conditions, namely, similar and dissimilar On the other hand, the reading span

task would comprise four conditions The participants would go through similar and dissimilar conditions which would comprise high context and low context sentences When being referred to as high context, it is implied that the sentences have been constructed in a form that that the last words

of the sentences could not be easily replaced by any other word and thus are high in contextual value The low context sentences are exactly the opposite and the last words of these sentences could

be easily replaced by any other word or was an unexpected word

words would be recalled

lar words would be better than that of the dissimilar words, a finding that would replicate Copeland and Radvansky (2001) In the extended part of the study, with high and low context

sentences used in the reading span tasks it was expected that the recall of the high context sentences would be greater than that of the low context sentences The sentences categorized as low sentencwould imply that the content of the sentences would be very low in contextual value as comparthe high context sentences which would be high in the contextual value It is suggested that the high context sentences would be easily comprehended and thus memorizing and recalling them It is expected that the high context sentences would be able to provide enough cues to help the participantrecall the words whereas it would not be the case in the low context sentences leading to a lower recall Critically, if sentence context is responsible for overcoming the phonological similarity decrement, the reversal of the traditional phonological similarity effect would be observed in the high context sentences but attenuated or abolished in the low context sentences

Chapter II: Pilot Study

Many studies investigating serial recall in short term memory adopted the procedure where

m and are also required to verbally call the words in the test (Copeland and Radvansky, 2001) However, there are other studies where

ll

e similarity effect, it seems prudent to determine if the

honological similarity effect found with simple word spans would be robust across different

words, 3 call

en

subsequent studies On the other hand, it is possible that the phonological similarity

participants read aloud the words as they are presented to the

re

participants silently read the words as they are presented on the screen and subsequently do either written recall (Bryan & Luszcz, 1999; Gustafson, 2001) on prepared answer sheets or verbal reca(Copeland and Radvansky, 2001)

Although the purpose of the present study is not to determine whether encoding and responsformats influence the phonological

condition required participants to say aloud the words during presentation and recall them verbally

If all the 3 conditions yielded a phonological similarity effect, then the logistically easier written recall procedure (verbal recall would require the running of single subjects) would be

adopted for the

effect would be strongest under conditions (verbal articulation) which reinforce phonological cod(Mueller, Seymour, Kieras, & Meyer, 2003).

Method

Participants

ing

105 undergraduate students from the Department of Social work and Psychology participated

dy for course credit 36 each were randomly assigned to the written and verbal conditions

There were 33 participants in the read and recall aloud condition

The condition variable was een subjects and composed of 3 levels, written, verbal and read and recall aloud

ondition and vice versa for the other half A short break was given in between these conditions

ing software 420 Words were selected in the form of 60 sets of seven words each Each

f seven words had the same rhyming scheme For example, bat, mat, cat; sink, wink, blink, e

in the stu

Design

A 3 (Condition) X 2 (List type) mixed design was adopted

completely betw

List type was run within subjects and comprised similar (rhyming) and dissimilar (non

rhyming) words Half the participants did the similar condition first followed by the dissimilar c

Materials

PCs were used to present the words to the participants The programs were designed by usthe E- Prime 1.1

of the set o

tc The words used in the lists presented to the participants were then checked for their

frequencies 10 sets of rhymes were put in each list Six lists were prepared for the study (See

Appendix A and Table 1) Each list was balanced for the average frequencies and standard

deviations of all the sets comprising it

Each list contained 10 sets of seven words each Each list was balanced for the number of

) or diphthongs (‘eI’, ‘@U’, c) and also for the rhymes used (example, rat, bat, mat; hut, but, cut; etc), a measure undertaken to ensure

ote: Frequencies are taken from the Kucera Francis (1967) counts

short vowels (‘a’, ‘e’, and ‘o’, etc) long vowels (‘a:’, ‘e:’, and ‘O:’ etc

et

that none of the word lists differed from each other in terms of phonetics and they were similar in terms of vowel length that could lead to word length effects (cf Baddeley, Thomson & Buchanan, 1975) A one-way between subjects analysis of variance (ANOVA) was done on these lists to ensure that there was no significant difference between the average means of the sets used in

the lists, F < 1 A balanced Latin – square procedure was used to rotate the lists across the 6

different conditions in the study Each participant therefore saw a different list for the similar and dissimilar conditions However, the Latin – square procedure ensured that each list was presented in each of the different conditions, but the different subjects

word at a ti

ed them to recall the words

All the participants were given four practice trials so as to ensure that the whole procedure was understood Upon assurance

e

own a sequence of words starting from a list length of 3 words per trial The participants were shown one word at a time for 1 second per word There were 2 trials at each list length The number of words shown was increased by one for the next list length up till 7 words were shown in each trial For similar words, the program controlling the experiment randomly sampled without replacement the appropriate number of words from each rhyming set (e.g., tame, same, lame, game, fame, name, came) For dissimilar words, one word was randomly sampled without replacement from a different rhyming set (e.g., tame, bet, creep, butt, lock, pill, sun) Phonological similarity was therefore manipulated within a trial Half the participants did the similar condition first followed by the dissimilar condition and vice versa for the other half A short break was given in between these

conditions None of the words presented to the participants in any one of the trials were repeated in subsequent trials in the whole experiment

In the written condition they had to silently read the words appearing on the screen and then

to recall the words and type them in the same order as they had seen If the participants were unable

to recall any particular word, they were asked to type “?” instead of the word at the particular position

ove to the next word After pressing the enter key, another box was be displayed in which they had to type the second

shown the words in written condition The only difference was in the recall phase when

and recall the rest In the verbal condition, the procedure was the same, except that they were asked to recall the words aloud instead of typing them Instructions given to the participants in the read and recall aloud condition were different from the ones given in the previous two conditions They were asked to read aloud as the words appeared on the screen and later when instructed to recall the words, they should do the same by recalling the words aloud

During the recall phase in the written condition the computer screen displayed a small box in which they had to type the first word of the sequence and then press enter key to m

word and so on After the participants recalled words after each trial, they had to press the enter key to continue to the next trial They were only allowed to make changes to the word they typed before they pressed the enter key The participants were not allowed to go back to the word they had typed after they had pressed the enter key Recall of all the participants was saved for analysis

For verbal condition, the participants were shown the words on the screen in the same way as they were