Rick's Home Page | Dissertation Table of Contents

CHAPTER 3
 
METHOD
 
This chapter contains an overview of the research design and defines the variables investigated by the study. A description of the subjects who participated in the study and an explanation of the instruments and procedures used in conducting the study follows. Finally, a description of the procedures used to analyze the data is given.
 
Overview of Research Design

The study was conducted in two rounds with the same subjects completing the two rounds sequentially. Each round of the study employed a pretest-posttest design (Brown, 1988, p. 154) with an experimental group and a control group. The pretest for each round served to control for vocabulary knowledge prior to the treatment. The treatment for each round included three conditions as follows: (1) an individual communication task introducing new lexical items to individual members of the experimental group, (2) a group communication task introducing new lexical items to small groups of members of the experimental group, and (3) unrelated activities for the control group. The posttest for each round measured both productive and receptive vocabulary knowledge, and served as a measure of acquisition (i.e., long-term recall) of the new lexical items introduced through the communication tasks conducted during the treatment. Please see Figure 1 for an overview of the research design.

Figure 1. Overview of Research Design

+---------------------------------------------------------------+
¦     ¦           ¦                                             ¦
¦  R  ¦  PRETEST  ¦  administered individually to all subjects  ¦
¦     ¦           ¦                                             ¦
¦  O  +-----------+---------------------------------------------¦
¦     ¦           ¦             ¦Task 1A conducted with         ¦
¦  U  ¦           ¦ Condition 1 ¦individual members of the      ¦
¦     ¦           ¦             ¦experimental group             ¦
¦  N  ¦           +-------------+-------------------------------¦
¦     ¦           ¦             ¦Task 1B conducted with small   ¦
¦  D  ¦ TREATMENT ¦ Condition 2 ¦groups of members of the       ¦
¦     ¦           ¦             ¦experimental group             ¦
¦     ¦           +-------------+-------------------------------¦
¦     ¦           ¦             ¦unrelated activities conducted ¦
¦  O  ¦           ¦ Condition 3 ¦with members of the control    ¦
¦     ¦           ¦             ¦group                          ¦
¦  N  +-----------+---------------------------------------------¦
¦     ¦           ¦                                             ¦
¦  E  ¦ POSTTEST  ¦ administered individually to all subjects   ¦
¦     ¦           ¦                                             ¦
¦-----+-----------+---------------------------------------------¦
¦     ¦           ¦                                             ¦
¦  R  ¦ PRETEST   ¦ administered individually to all subjects   ¦
¦     ¦           ¦                                             ¦
¦  O  +-----------+---------------------------------------------¦
¦     ¦           ¦             ¦Task 2A conducted with         ¦
¦  U  ¦           ¦ Condition 1 ¦individual members of the      ¦
¦     ¦           ¦             ¦experimental group             ¦
¦  N  ¦           +-------------+-------------------------------¦
¦     ¦           ¦             ¦Task 2B conducted with small   ¦
¦  D  ¦ TREATMENT ¦ Condition 2 ¦groups of members of the       ¦
¦     ¦           ¦             ¦experimental group             ¦
¦     ¦           +-------------+-------------------------------¦
¦     ¦           ¦             ¦unrelated activities conducted ¦
¦  T  ¦           ¦ Condition 3 ¦with members of the control    ¦
¦     ¦           ¦             ¦group                          ¦
¦  W  +-----------+---------------------------------------------¦
¦     ¦           ¦                                             ¦
¦  O  ¦ POSTTEST  ¦ administered individually to all subjects   ¦
¦     ¦           ¦                                             ¦
+---------------------------------------------------------------+

 
Variables

The main variables which the present study investigates are interaction, comprehension, and acquisition of new lexical items. The extent to which these variables occur in individual settings as compared to group settings is also investigated. Additionally, the degree of overall lexical acquisition attained by the experimental group is compared with the degree of overall lexical acquisition attained by the control group.

As reflected in the research questions, the present study includes an investigation of the relationships among these variables (i.e, interaction, comprehension, and lexical acquisition). In the group setting, the study also allows for the investigation of the relationships among these variables for the subjects who actively participate in the interaction as compared to those subjects who are merely exposed to the interaction.

Hence, a complete list of variables investigated by the present study is as follows: (1) interaction, (2) comprehension, (3) lexical acquisition, (4) setting, which has two levels: individual and group, (5) group, which has two levels: experimental and control, and (6) participation, which has two levels: active and passive.

 
Subjects

The identification of subjects for this study took place at the beginning of the school-year (September, 1991) at an American elementary school located in a self-contained American community in Japan. The school's population consists of approximately 1,000 students in kindergarten through sixth grade; over the last 10 years there has been an average of 45 students enrolled in the school's ESL program. The subjects for the study consisted of the 14 kindergarten students who had been determined to be eligible for the ESL program. These students attend a regular kindergarten session of 2½ hours with native-English-speaking peers and teachers; additionally, they attend a daily ESL class of 45 minutes with one of two native-English-speaking ESL teachers. At the beginning of the study, the subjects' ages ranged from 5 years and 2 months to 6 years and 2 months and represented three first-language backgrounds: Japanese, Tagalog, and Thai. Please see Appendix A for additional background information about each subject.

The 14 subjects were randomly assigned to either the experimental group or the control group and remained in their assigned group for both rounds of the study. Three of the subjects were selected to serve as the control group, with 11 forming the experimental group; illness and absence affected the experimental group. Illness of 1 student during the first round left 10 students in the experimental group; that student participated in the second round. However, the absence of another student again left 10 students in the experimental group for Round 2. The control group remained intact for both rounds of the study.
 

Instruments and Procedures
 
Placement in ESL Classes

Eligibility for the ESL program is based on results of the PRE-LAS (Language Assessment Scales) (Duncan & De Avila, 1985) administered at the beginning of the school-year. Scoring on the PRE-LAS provides for five levels of English proficiency, Level 1 being a non-English-speaker and Level 5 being a nativelike English speaker. Students scoring in PRE-LAS Levels 1 through 3 are considered eligible for the ESL program; the results played no other role in the study. The PRE-LAS is an untimed test, administered individually; it was administered and scored by one of two ESL teachers who followed the procedures as described in the instruction manual. Results of the PRE-LAS are shown in Appendix B.
 

Lexical Proficiency

The Peabody Picture Vocabulary Test (PPVT) (Dunn & Dunn, 1981) is designed primarily to measure a subject's receptive vocabulary in Standard American English. The PPVT was used as a measure of general lexical proficiency, as the study investigates vocabulary development. The PPVT is an untimed test, administered individually; it was administered and scored by the researcher who followed the procedures as described in the instruction manual. Results of the PPVT are shown in Appendix B.
 

Pretests
 
Content

The pretest for the first round consisted of words from the following categories: colors, bugs, and birds. The words are vocabulary items from the Peabody Language Development Kit (Dunn & Smith, 1965) and are part of the ESL curriculum. At the time of the study, colors had been taught, whereas bugs and birds had not yet been introduced in class.

The pretest for the second round consisted of words from the following categories: fruit, vegetables, and occupations. The fruit-names and vegetable-names are vocabulary items found in Lexicarry (Moran, 1984) and are part of the ESL curriculum, while the occupation-names are vocabulary items from the Peabody Language Development Kit (Dunn & Smith, 1965) and are also part of the ESL curriculum. At the time of the study, none of these words had been introduced in class.
 
Aim
 
The pretests for both rounds measured both productive and receptive vocabulary knowledge, and served to control for vocabulary knowledge prior to the treatment phase of the study.
 
Administration

The pretests for both rounds were administered individually by the researcher. Productive vocabulary knowledge was measured by the subject's ability to verbalize the names of flash-cards representing the items. The researcher asked the subject to name each flash-card while displaying them in turn, allowing ample time for the subject to respond. Each set of flash-cards was presented separately. In Round 1, first colors; then bugs; and finally, birds were presented. In Round 2, first fruit; then vegetables; and finally, occupations were presented.

The pretest for receptive vocabulary knowledge immediately followed the pretest for productive vocabulary knowledge. Employing a forced-choice format, the researcher asked the subject to point to one item on each page of a picture booklet, allowing ample time for the subject to respond. Each page contained six pictures (i.e., the correct picture and five distracters). Receptive vocabulary knowledge was measured by the subject's ability to identify the appropriate picture. Three picture booklets, one for each set of words, were used. Only pictures representing words from within each set were used in each picture booklet. Each picture booklet was presented separately. In Round 1, first colors; then bugs; and finally, birds were presented. In Round 2, first fruit; then vegetables; and finally, occupations were presented.

All words introduced as new lexical items in the communication tasks during the treatment phase (i.e., bugs, and birds in Round 1; and fruits, vegetables, and occupations in Round 2) were concrete nouns. Please see Appendix C for a complete list of lexical items included in the pretest.

 
Selection of Lexical Items for Communication Tasks
 
Round 1

Based on the results of the pretest in the first round, the researcher selected vocabulary items for use in the treatment phase of Round 1.

Six color-names that were known receptively by all subjects were selected for use in both the individual and group tasks. As colors were part of the teacher's directions in the communication tasks and not introduced as new lexical items, it was necessary to make certain that these color-names were known by the subjects.

Flash-cards representing four bug-names and four bird-names which were unknown productively and receptively to each subject in the experimental group were selected for use in the individual communication task in Round 1. Three of these four bug-names and three of the four bird-names were explicitly included in the teacher's instructions while the remaining two flash-cards (one bug and one bird) served as distracters.

Flash-cards representing four bug-names and four bird-names which were unknown productively and receptively to all subjects in each subgroup of the experimental group were selected for use in the group communication task in Round 1. These bug- and bird-names were different from those used in the individual task. Three of the four bug-names and three of the four bird-names were explicitly included in the teacher's instructions while the remaining two flash-cards (one bug and one bird) served as distracters.

Thus six words were introduced as new lexical items in both the individual and group communication tasks during the first round.
 
Round 2

Based on the results of the pretest in the second round, the researcher selected vocabulary items for use in the treatment phase of Round 2.

Flash-cards representing four fruit-names, four vegetable-names, and three occupation-names which were unknown productively and receptively to each subject in the experimental group were selected for use in the individual communication task in Round 2. Three of these four fruit-names and three of the four vegetable-names were explicitly included in the teacher's instructions while the remaining two flash-cards (one fruit and one vegetable) served as distracters. All three of the occupation-names were explicitly included in the teacher's instructions.

Flash-cards representing four fruit-names, four vegetable-names, and three occupation-names which were unknown productively and receptively to all subjects in each subgroup of the experimental group were selected for use in the group communication task in Round 2. These fruit-, vegetable-, and occupation-names were different from those used in the individual task. Three of the four fruit-names and three of the four vegetable-names were explicitly included in the teacher's instructions while the remaining two flash-cards (one fruit and one vegetable) served as distracters. All three of the occupation-names were explicitly stated in the teacher's instructions on two separate occasions.

Thus nine words were introduced as new lexical items in both the individual and group communication tasks during the second round.
 

Treatment

The treatment in both rounds of the study consisted of three conditions; however, every attempt was made to conduct the treatment so that it would appear to the subjects to be a regular class activity.
 
Condition 1
 
Condition 1 was met by each member of the experimental group participating in an individual communication task conducted by the researcher. Both the individual communication task in the first round of the study (Task 1A), and the individual communication task in the second round of the study (Task 2A) introduced new lexical items to these subjects.
 
Condition 2
 
Condition 2 was met by two subgroups of the experimental group participating in a group communication task conducted by the researcher. The experimental group of 10 students was divided into two subgroups of 5 students each for the group task in each round. There was no difference in the conditions for these two subgroups as it is the norm at the data collection site for ESL classes to meet in small groups. Both the group communication task in the first round of the study (Task 1B), and the group communication task in the second round of the study (Task 2B) introduced new lexical items to these subjects.
 
Condition 3
 
Condition 3 was met by the control group participating in unrelated activities with another teacher during the time the members of the experimental group participated in the communication tasks. This was accomplished by arranging for the control group to visit the computer center at the designated time of the group communication task.

 
The Communication Tasks

As in previous research in this area carried out by other researchers, this study used communication tasks to serve as the vehicle for data collection. Of the three communication tasks (jig-saw task, information-gap task, and opinion-exchange task) described by Pica (1992), it is the information-gap task that has proved to be most suitable for very young learners. In classroom activities involving communication tasks conducted with kindergartners at the data-collection site of this study, the researcher and a colleague both found the opinion-gap tasks to be beyond the linguistic, or perhaps cognitive skills of the subjects, while jig-saw tasks were beyond their social skills. Furthermore, information-gap tasks allow the teacher to hold the crucial, task-relevant information through which the new lexical items are introduced. Considering these advantages, four information-gap tasks were developed by the researcher for use in the study.
 
Content

All four of these tasks involved the subjects' selecting certain flash-cards and placing them at a particular destination on a board in front of them, as instructed by the teacher. Embedded in the teacher's instructions were lexical items which were unknown to the subjects. As it was established before hand that the new lexical items were in fact unknown to the subjects, it was predicted that the subjects would request the information held by the teacher (i.e., the word meanings) in order to successfully complete the task. The tasks themselves were simple enough that the researcher believed that the only obstacle for the subjects to complete the task would be the meanings of the new lexical items. A pilot study conducted the previous school-year confirmed these predictions.
 
Task 1A (Individual Zoo Keeper). This task served as the individual information-gap task for Round 1. The names of the flash-cards that the subjects were instructed to select were new lexical items. Please see Figure 2 for a diagram representing Task 1A..ls 1
 
Task 1B (Group Zoo Keeper). This task served as the group information-gap task for Round 1. The names of the flash-cards that the subjects were instructed to select were new lexical items. Please see Figure 3 for a diagram representing Task 1B.
 
Task 2A (Individual Snack Time). This task served as the individual information-gap task for Round 2. The names of the flash-cards that the subjects were instructed to select were new lexical items. The names of the destinations at which the subjects were instructed to place these flash-cards were also new lexical items. Please see Figure 4 for a diagram representing Task 2A.
 
Task 2B (Group Snack Time). This task served as the group information-gap task for Round 2. The names of the flash-cards that the subjects were instructed to select were new lexical items. The names of the destinations at which the subjects were instructed to place these flash-cards were also new lexical items. Please see Figure 5 for a diagram representing Task 2B.

--------------

Figure 2

--------------

Figure 3

--------------

Figure 4

--------------

Figure 5

--------------
 
Aim

The individual tasks served to introduce new lexical items to individual subjects; stimulate interaction between the researcher and the subjects; and to measure the subjects' comprehension of the input containing the new lexical items, as measured by the level of success achieved in completing the tasks. Analyzing transcriptions of the interaction which occurred during the individual tasks, allowed the researcher to investigate the effect of negotiated interaction on the comprehension of input containing new lexical items.

The group tasks served to introduce new lexical items to small groups of subjects; elicit interaction between the researcher and the subjects and possibly among the subjects; and to measure the subjects' comprehension of the input containing the new lexical items, as measured by the level of success the subjects achieved in completing the tasks. Subjects were free to interact, or not to interact, at their discretion. Thus, analyzing transcriptions of the interaction which occurred during the group tasks, allows the researcher to investigate the effect that participation in the interaction has on the comprehension of input containing new lexical items as compared to the effect that exposure to the interaction has on the comprehension of input containing new lexical items.
 
Administration

For a detailed description of the tasks, please see Appendix D. Additionally, Appendix E contains complete transcriptions of the information-gap tasks. General guidelines followed by the researcher when carrying out all information-gap tasks are as follows:
 

Posttests

Both rounds of the study employed a posttest which was administered individually by the researcher 7 days after completion of the tasks. The posttest for each round was identical to the pretest for that round; it contained all words contained in the pretest and employed the same flash-card format to measure productive vocabulary knowledge and the same forced-choice, picture booklet format to measure receptive vocabulary knowledge as that used in the pretest. Please see Appendix C for a complete list of lexical items included in the posttest. The posttest served as an indication of lexical acquisition to the extent that it measured long-term recall of the words.

Scoring

To investigate the research questions posed by the present study, the three variables (interaction, comprehension, and lexical acquisition) were operationalized and a scoring system for each was developed. This section describes the procedures used in scoring the data collected. The interaction and comprehension scores are derived from analysis of the treatment phases of the study (i.e., the communication tasks) while lexical acquisition scores are derived from results of the pre- and posttests.

Interaction

The type of interaction most relevant to the present study is verbal interaction which contains negotiation for the meaning of a new lexical item, as it is theorized that this negotiation brings about the interactional adjustments which makes the input comprehensible to the learners and plays an essential role in the acquisition process. Varonis and Gass focused on non-native-speaker discourse to present their model for the negotiation of meaning or "non-understanding routines" (1985b, p. 73) which they define as "those exchanges in which there is some overt indication that understanding between participants has not been complete" (1985b, p. 73). Please see Figure 6 for a diagram of this model which Varonis and Gass describe as follows:

The first part of the model consists of a trigger, denoted by T. The second part of the model we term the resolution. It consists of an indicator (which we denote by I), a response (denoted by R), and a reaction to the response (denoted by RR). Simply put, the trigger is that utterance on the part of the speaker which results in some indication of non-understanding on the part of the hearer. (1985b, p. 74) Figure 6. Model for Non-understandings
          (from Varonis and Gass 1985b, p. 74)

----------------------------------------------------------

          Trigger                  Resolution

           +---+                 +------------+
           ¦ T ¦        »        ¦ I » R » RR ¦
           +---+                 +------------+

----------------------------------------------------------
 
Two refinements to this model described in subsequent studies are relevant to the present study. First, to allow for a more detailed analysis of the data, the indicator is subdivided into two classifications, namely, direct and indirect indicators of non-understanding (Gass & Varonis, 1985a). Gass and Varonis (1985a) explain that direct indicators leave no doubt that there has been a lack of understanding and are often expressed with a wh- word, while indirect indicators are a more gentle means of expressing a lack of understanding and are often a repetition of a word or phrase from the previous utterance. The present study recognizes both direct and indirect indicators of non-understanding under the condition that the indicator is verbal; that is, for scoring purposes, silence is not considered to be an indicator of non-understanding. Second, Gass and Varonis define negotiation to be "those exchanges in which there is some overt indication that understanding between participants has not been complete and there is a resultant attempt to clarify the non-understanding" (1985b, p. 39). The present study considers this attempt to clarify the non-understanding to be met if the teacher's response to the learner's indicator contains additional information. That is, for scoring purposes in the present study, a non-understanding routine is considered complete when the teacher supplies additional information in response to a learner's utterance which signals, directly or indirectly, the non-understanding of a target word.

After the tasks were recorded and transcribed, in order to quantify the interaction which occurred, the transcriptions were scored in terms of the model for non-understandings presented by Varonis and Gass (1985b) with consideration given to the two refinements described above. The flow chart used as the model for scoring negotiated interaction in the present study is displayed in Figure 7.

--------------

Figure 7

--------------

This model does not assume, however, that understanding is attained, only that negotiated interaction has occurred. Neither does it preclude the possibility of a single target word resulting in numerous non-understanding routines.

As a measure of interaction, the present study assigned 1 point to the subjects for each non-understanding routine in which the trigger (T) was a new lexical item introduced in the task, the indicator (I) was a subject's utterance(s) which overtly marked non-understanding of the trigger, and the response (R) was the teacher's utterance(s) which provided additional information to the subject. The present study considers the reaction to the response (RR) an optional unit of the routine, as do Varonis and Gass (1985b), thus its absence or presence plays no role in scoring the interaction. An example of a complete non-understanding routine taken from the data collected for the present study appears here in Excerpt 1.

Excerpt 1.
---------------------------------------------------------
Teacher: please give the papaya to the roofer/       (T)

Subject: what's the roofer? like this?/              (I)

Teacher: a roofer is a person who climbs up          (R)
         on top of the house to fix the roof/

Subject: okay/                                       (RR)
----------------------------------------------------------
 
The subject in the above negotiated interaction (Excerpt 1) received 1 interaction point for having brought about this interactional adjustment by first noticing the trigger "roofer," indicating non-understanding by asking for a definition of the unknown word which, in turn, brought about a response by the teacher which provided additional information. The subject's subsequent reaction to the teacher's response ("okay") played no role in the scoring. That is, for the subject to receive an interaction point the trigger (a new lexical item), the indicator (a verbal indication of non-understanding), and the response (providing new information) must all be present.

In Excerpt 1, the indicator of non-understanding is a direct indicator, as the subject specifically asks for information about the target word. An example of an indirect indicator of non-understanding appears here in Excerpt 2.

Excerpt 2.
----------------------------------------------------------
Teacher: are you ready for the next one?/ okay/ +
         please find the coconut=                     (T)

Subject: =coconut/=                                   (I)

Teacher: =and give it to the veterinarian/

Subject: coconut/                                     (I)

Teacher: do you know what a coconut is?/ + a
         coconut is like a big brown ball and it      (R)
         looks like it has hair on it/ ++
----------------------------------------------------------

In this case, the subject received 1 interaction point for having brought about this interactional adjustment by first noticing the trigger "coconut," indicating non-understanding with an "echo" (Gass & Varonis, 1986, p. 334) by repeating part of the previous utterance in a falling intonation pattern which, in turn, brought about a response by the teacher which provided additional information. The subject in fact produced two indirect indicators but received only 1 point, as only one of the indicators brought about a response by the teacher.

Both of the excerpts above are rather simple exchanges. An example of a more complex exchange appears here in Excerpt 3.
 
Excerpt 3.
----------------------------------------------------------
Teacher: please find the spinach and=                 (T)

Subject: =what it look like?/=                        (I)

Teacher: =give it to the veterinarian/ okay/
         spinach looks like green leaves on           (R)
         little stems/ big green leaves/

Subject: ah look like corn?/                          (I)

Teacher: well it doesn't look like corn/              (R)
         corn is yellow/ spinach is green/
         green leaf/ green/ it's green/
         with a lot of leaves/ big leaves/ ++

Subject: i got it/ ++ i got it already/               (RR)
----------------------------------------------------------

The subject in the above negotiated interaction (Excerpt 3) received 2 interaction points as both of the indicators brought about responses by the teacher which provided additional information. The new lexical item "spinach" provides the trigger for both indicators.

Of course it is conceivable that a subject may attempt to carry out the instructions without negotiating for the meaning of an unknown lexical item, as interaction is not the only option available to learners who encounter new linguistic material. Færch and Kasper point out that "inference" (1986, p. 270) can be a useful tool for learners, and Sharwood Smith (1986) contends that other knowledge systems (e.g., the L1 system, gesture, logical inferences, the current situation) may be accessed to make input comprehensible. Considering the fact that in the tasks that comprise the treatment phase of the present study, there is a finite number of flash-cards to select from and places to put them, it is certainly within the realm of possibility that the learners could carry out the instructions without negotiating and do so correctly.
 

Comprehension

Comprehension was measured by the subjects' success in completing the tasks as directed by the teacher. Subjects received 1 point for the correct selection of a flash-card and 1 point for the correct placement of the flash-card. Subjects were directed to select and place six flash-cards for a total of 12 possible points. In Round 1 of the study, the names of the six flash-cards to be selected were new lexical items; however, the names of the locations at which they were to be placed were known. In Round 2, the six names of the flash-cards to be selected and the three names of the locations at which they were to be placed were new lexical items. At appropriate times throughout the execution of each task, the researcher checked each subject's board and marked a score sheet accordingly. Scores were tallied after completion of the tasks.
 

Lexical Acquisition

As mentioned in Chapter 2, several views of vocabulary knowledge propose that the lexical store be viewed as a continuum rather than a dichotomy (e.g., Gass, 1988a; & Teichroew, 1982). According to Teichroew, "the boundaries of receptive knowledge and productive knowledge are variable" (1982, p. 21); therefore, for the purposes of the present study, the notions of productive and receptive vocabulary knowledge are tightly defined. A subject displays productive knowledge of a word by correctly verbalizing the names of flash-cards representing the items when asked to do so by the researcher. This flash-card format test is administered individually according to the procedures described above. A subject displays receptive knowledge of a word by correctly identifying the appropriate picture in a forced-choice test (described above) which is administered individually following similar procedures to that of the PPVT.

Lexical acquisition was measured for both productive and receptive knowledge of the new lexical items presented by administering the posttests 7 days after completion of the tasks. Round 1 of the study introduced six new lexical items to the subjects in each task (i.e., Tasks 1A and 1B), while Round 2 introduced nine new lexical items to the subjects in each task (i.e., Tasks 2A and 2B). For the productive lexical acquisition score, subjects received 1 point for each word used in the tasks and produced correctly in the posttest which had not been produced correctly in the pretest. For the receptive lexical acquisition score, subjects received 1 point for each word used in the tasks and identified correctly in the posttest which had not been identified correctly in the pretest.

In addition to serving as a tool for measuring vocabulary growth in the new lexical items presented in the tasks as described above, the overall pre- and posttest scores achieved by the experimental group are compared with those achieved by the control group (who did not participate in the treatment phase of the study) in order to establish whether there were any learning effects from completing the tests.
 

Reliability of Measures

Since the results of this study are based in part on quantitative measures of the variables (interaction, comprehension, and lexical acquisition), it is important to ensure that the procedures and tests used are consistent and accurate. This section describes the steps taken to address the issue of reliability of the quantitative measures used.
 

Interaction

A second rater was trained in the procedures for scoring the interaction as described above using the data from the pilot study. Then two subjects from each task were selected randomly for comparison with the scoring done by the researcher in order to test the reliability of the scoring method and to report a measure of the degree to which the scoring was done accurately and consistently. The trial-by-trial agreement (Page & Iwata, 1986) was calculated with the interaction resulting from the introduction of one new lexical item comprising one trial. This calculation yielded an inter-rater reliability index of 86.66%.
 

Comprehension

The measure of comprehension is simple and accurate. As described above, the researcher marks a score sheet indicating whether or not the subject has carried out the instructions correctly. However, since no scoring system is completely infallible, a safeguard was taken to double check the scoring. At the end of each session, the researcher called out the name and placement of each of the items on each student's board. Thus, by reviewing the recordings of the tasks, the researcher was able to make certain all score sheets were marked correctly. Therefore, it is safe to assume that the reliability of the measure of comprehension consistently approaches 100% accuracy.
 

Lexical Acquisition

Measures of the internal reliability of the tests of receptive vocabulary knowledge,3 used as the pre- and posttests, appear in Table 1.

 

Table 1. Reliability of Tests of Receptive Vocabulary
         Knowledge
-----------------------------------------------------------
Test                                             K-R20
-----------------------------------------------------------
Round 1 Pretest                                   .37

Round 1 Posttest                                  .76

Round 2 Pretest                                   .80

Round 2 Posttest                                  .82
-----------------------------------------------------------
 
All tests, with the exception of the pretest in Round 1, exhibit an acceptable level of reliability. It must be pointed out, however, that the low measure found in the Round 1 pretest appears to be more a problem of the skewed distribution of the data (all subjects scored relatively low on the test) than with actual test reliability. The reliability measures were calculated with the Kuder-Richardson formula 20 (K-R20) using the Statistical Package for the Social Sciences computer software (SPSS/PC+) (Norusis, 1990).
 

Analysis

The analysis begins with an overview of the results which includes a close look at the quantifications made for interaction, comprehension, and acquisition of the new lexical items introduced during the treatment phase of the present study. Additionally, the descriptive statistics of these measures are reviewed.

The analysis of interaction, comprehension, and lexical acquisition includes an investigation of the effect that the setting (individual or group) has on these measures. This analysis treats interaction, comprehension, and lexical acquisition as dependent variables with setting (individual or group) as the independent variable. This analysis was accomplished by calculating Multivariate analysis of variance procedures and Univariate F-tests using SPSS/PC+ (Norusis, 1990). The alpha level for statistical decisions based on these calculations was set at .05.

Additionally, the degree of overall lexical acquisition attained by the experimental group (those subjects who participated in the tasks) is compared with the degree of lexical acquisition attained by the control group (those who did not participate in the tasks). Overall lexical acquisition is measured by comparing total pre- and posttest scores for the members of each group (experimental and control). The analysis was accomplished by calculating Group by Test Repeated-measures Analysis of Variance (ANOVA), using SPSS/PC+ (Norusis, 1990), for each round of the study. The alpha level for statistical decisions based on these calculations was set at .05.

Further investigation is intended to shed light on the issues addressed by the research questions presented in Chapter 1. Each research question, along with an explanation of the types of analysis used to accomplish its investigation, is presented here.
 

Research Question 1

What is the relationship between interaction and the comprehension of input containing new lexical items?

The analysis of this question entails an examination of the descriptive statistics and the comparisons made between settings for the variables addressed (i.e., comprehension and lexical acquisition). Additionally, correlations between the interaction scores and the comprehension scores using the Pearson product moment correlation coefficient were calculated, using SPSS/PC+ (Norusis, 1990), to determine the degree of relationship between these variables. Due to the fact that several correlations were calculated from scores achieved by the same subjects, the alpha level for statistical decisions of .05 was divided in half and set at a more conservative .025 for all correlations.
 

Research Question 2

What is the relationship between interaction and the acquisition of these new lexical items?

The analysis of this question entails an examination of the descriptive statistics and the comparisons made between settings for the variables addressed (i.e., interaction and lexical acquisition). Additionally, correlations between the interaction scores and the acquisition scores using the Pearson product moment correlation coefficient were calculated to determine the degree of relationship between these variables. A decision was made post hoc to calculate two additional correlations (i.e., rehearsal/acquisition and teacher repetition/acquisition, described below). Due to the fact that several correlations were calculated from scores achieved by the same subjects, the alpha level for statistical decisions of .05 was divided in half and set at a more conservative .025 for all correlations. Correlation coefficients were calculated using SPSS/PC+ (Norusis, 1990).
 

Research Question 3

What is the effect of the comprehension of input containing new lexical items on the acquisition of these new lexical items?

The analysis of this question entails an examination of the descriptive statistics and the comparisons made between settings for the variables addressed (i.e., comprehension and lexical acquisition). Additionally, correlations between the comprehension scores and the acquisition scores using the Pearson product moment correlation coefficient were calculated, using SPSS/PC+ (Norusis, 1990), to determine the degree of relationship between these variables. Due to the fact that several correlations were calculated from scores achieved by the same subjects, the alpha level for statistical decisions of .05 was divided in half and set at a more conservative .025 for all correlations.
 

Research Question 4

During group interaction, do the relationships among interaction, comprehension, and acquisition differ for those subjects who orally participate in the interaction as compared to those subjects who are exposed to the interaction but choose not to orally participate?

The analysis of this question entails a review of the descriptive statistics of the three variables and the correlations calculated for Research Questions 1 through 3. Additionally, to investigate the effect that participation has on the relationships among the variables, the subjects who completed the group information-gap tasks in each round are subdivided into classifications based on the amount of negotiated interaction they initiate. Fisher Exact probability tests (Siegel, 1956, p. 96) are calculated to determine if those subjects who orally participate in the interaction achieve comprehension and acquisition of the new lexical items differently than those subjects who are exposed to the interaction but choose not to orally participate. The alpha level for statistical decisions based on these calculations was set at .05.
 

Summary

This section summarizes the procedures of the present study. Kindergarten ESL students (N=14) are administered a pretest to determine unknown lexical items. An experimental group (n=11) then participates in information-gap tasks with their ESL teacher which are similar to class activities carried out on a regular basis; however, these particular tasks explicitly include unknown lexical items in the teacher's instructions. A control group (n=3) is administered the pretest, but does not participate in the tasks.

The study is conducted in two similar rounds, each consisting of individual tasks and group tasks. There are, however, differences between the two rounds. Round 1 introduces six new words to the subjects, while Round 2 introduces nine new words. Also, in Round 1 the teacher gives feedback to the subjects only after their verbal indication of non-understanding, while in Round 2 the teacher gives unsolicited feedback to the subjects as to whether or not their selection and placement of the items is correct.

The interaction which occurs during the tasks is quantified by employing a model of non-understanding presented by Varonis and Gass (1985b). Comprehension is measured by the subjects' success in completing the tasks. Lexical acquisition is measured by a posttest administered to both the experimental group and the control group 7 days after the experimental group completes the tasks.

The results of these measures are analyzed in order to shed light on the issues addressed by the research questions presented in Chapter 1.


Rick's Home Page | Dissertation Table of Contents
 © by Rick Heimbach 1993 All Rights Reserved