Difference between revisions of "French gender attention"

From LearnLab
Jump to: navigation, search
(New page: {| class="wikitable" border="1" style="margin: 2em auto 2em auto" |- ! PIs | Presson, MacWhinney |- ! Faculty | MacWhinney |- ! Postdocs | Pavlik |- ! Others with > 160 hours | n/a |- ! ...)
 
(Explanation)
 
(18 intermediate revisions by 2 users not shown)
Line 14: Line 14:
 
|-
 
|-
 
! Study Start Date
 
! Study Start Date
| 03/06/08
+
| 10/01/08
 
|-
 
|-
 
! Study End Date
 
! Study End Date
| 06/30/08
+
| 05/30/08
 
|-
 
|-
 
! Learnlab
 
! Learnlab
Line 23: Line 23:
 
|-
 
|-
 
! Number of participants (total)
 
! Number of participants (total)
| 45
+
| 80-100
 
|-
 
|-
 
! Number of participants (treatment)
 
! Number of participants (treatment)
| 45
+
| N/A
 
|-
 
|-
 
! Total Participant Hours
 
! Total Participant Hours
| ~135
+
| ~100
 
|-
 
|-
 
! Datashop?
 
! Datashop?
| '08
+
| Summer '09
 
|}
 
|}
  
 
==Abstract==
 
==Abstract==
  
This study extends the research question in a previous study by [[French gender prototypes| Presson & MacWhinney]].  In that result, explicit instruction led to better generalization and robustness to forgetting, compared to correct/incorrect feedback only.  In the current study, explicitness is manipulated as well, with the more explicit condition the same as the previous study.  However, the more implicit condition in this case is a highlighting manipulation, with no rules presented and in which participants see the relevant cue in capital letters to draw attention to the endings of words (e.g. "fromAGE" v. "fromage").  We want to know first if students are as effective at extracting cue patterns with this intervention as with strictly explicit instruction.
+
This study extends the research question in a previous study by [[French gender prototypes| Presson & MacWhinney]].  In that experiment, explicit instruction led to better generalization and robustness to forgetting, compared to correct/incorrect feedback only.  In the current study, explicitness is manipulated as well, with the more explicit condition the same as the previous study.  However, the more implicit condition in this case is a highlighting manipulation, with no rules presented and in which participants see the relevant cue in capital letters to draw attention to the endings of words (e.g. "fromAGE" v. "fromage").  We want to know first if students are as effective at extracting cue patterns with this intervention as with strictly explicit instruction.
  
Second, we manipulate the time pressure of the task between groups.  Time pressure favors procedural and automatic performance. If the highlighting condition is less explicit and equally effective, we expect an advantage within the time pressure condition.  This is important because time pressure is an essential element of naturalistic language use.  
+
Second, we manipulate the time pressure of the task between groups.  Time pressure favors procedural and automatic performance. If the highlighting condition is less explicit and equally effective, we expect an advantage within the time pressure condition (answer within 1400ms) compared to the no time pressure condition (trial times out at 6000ms, as in prior studies).  This is important because time pressure is an essential element of naturalistic language use. In addition, time pressure changes the task demands, which we hypothesize will encourage more implicit representations of categories than with no time pressure.
  
 
==Glossary==
 
==Glossary==
Line 60: Line 60:
  
  
We are changing instruction here based on whether or not the relevant cue is presented with a salient (because of its frequency and simplicity) prototype or with a larger number of varied exemplars. 
+
In the current study, we manipulate whether participants receive explicit cues to gender patterns, or merely correct / incorrect feedback with the relevant cues highlighted (through capitalization) to direct attention.
 
 
In the current study, we also manipulate whether participants receive explicit cues to gender patterns, or merely correct / incorrect feedback.  Half of all cues are presented explicitly, half without explicit prompts (counterbalanced and randomized for each participant).
 
  
 
==Dependent variables==
 
==Dependent variables==
Line 70: Line 68:
 
==Independent variables==
 
==Independent variables==
  
Between-groups manipulations of time pressure and explicitness produce four groups:
+
Between-groups manipulations of time pressure (present or absent 1400ms time limit) and explicitness produce four groups:
  
 
Rule + Time Pressure
 
Rule + Time Pressure
 +
 
Rule - Time Pressure
 
Rule - Time Pressure
 +
 
Highlighting + Time Pressure
 
Highlighting + Time Pressure
 +
 
Highlighting - Time Pressure
 
Highlighting - Time Pressure
  
Time pressure is operationalized by timimg out test trials after (''some interval'') ms.  
+
Time pressure is operationalized by timimg out test trials after (''some interval'') ms.
 +
 
 +
 
 +
==Screenshots==
 +
 
 +
<gallery>
 +
Image:Gender pretest.jpg|Pre-test trial
 +
Image:Gender explicit.jpg|Explicit condition
 +
Image:Gender highlight.jpg|Highlighting condition
 +
Image:Gender feedback.jpg |Feedback
 +
 
 +
</gallery>
  
 
==Hypotheses==
 
==Hypotheses==
PENDING
+
After 40 minutes of training, we hypothesize learning that is retained after a one-week retention interval and which generalizes to novel words under the same cue categories. 
 +
 
 +
From previous studies, the response form has changed from seeing nouns in isolation and categorizing them as masculine or feminine to seeing both the correct and incorrect noun phrases on each trial, including a gendered definite article (e.g. "le fromage" and *"la fromage").  This method change may make learning more difficult than prior studies; however, it improves external validity by making the categorization process better resemble language input.
 +
 
 +
Because prior studies (e.g., [[French gender prototypes|Explicitness and Category Breadth]]) suggest that explicitness of instruction does not affect accuracy in training, we predict equivalent performance in the learning block.  Explicit presentation of the prompt would be more robust to decay and generalization if in fact the manipulation makes the cue more salient (compared to the highlighting condition).
 +
 
 +
Adding time pressure will suppress initial performance in training.  We predict that this suppression will be temporary, and in fact that later performance could show a time pressure advantage, especially in the no-feedback post-test.  In addition, time pressure should change the relative effectiveness of an explicit inference rule vs. noticing highlights, such that noticing is easier than inference under time pressure, therefore improving learning.
 +
 
 +
At the one-week delay, participants complete an auditory version of the categorization task, testing whether the cue information transfers at all to correct French phonological representation, is too distant from French to help with auditory categorization, or is purely orthographic / visual.
  
 
==Findings==
 
==Findings==
 +
Analyses were performed using d-prime as a measure of categorization sensitivity that accounts for possible response bias, and using response latency for correct responses.
 +
 +
#Pre-test performance was close to chance for all groups (mean accuracy = .51, ''SD''= .11; highest group mean d-prime = 0.30, ''SD''= 0.19).
 +
#  All groups learned the cues to grammatical gender as demonstrated by improved post-test d-prime and reduced post-test latency, both of which remained at a one-week delayed post-test.
 +
#  Explicit instruction showed higher accuracy than a less explicit highlighting condition, on tests both with time pressure (d-prime of explicit = 2.04, highlighting = 1.45) and without (d-prime of explicit = 2.70, highlighting = 1.96).  This contradicts our speculation that time pressure might mitigate the positive effect
 +
#  Both explicit and highlighting instructional groups had lower d-prime when tested under time pressure.
 +
#  Training under time pressure did not eliminate the impairment in accuracy from testing under time pressure (i.e., both groups had higher d-prime without time pressure).
 +
#  Training under time pressure did not affect d-prime at any post-test.
 +
#  After controlling for mean pre-test latency, neither instructional variable (explicitness or training time pressure) influenced the amount of speed-up from pre-test to post-test.  However, there was a marginally significant interaction (''F''(1,76) = 3.16, ''p'' = .08) such that the time pressure training responded faster at immediate post-test and that advantage disappeared after a one-week delay.
  
PENDING
+
==Explanation==
  
 +
Participants in all groups learned to categorize nouns by grammatical gender based on spelling cues. However, those who trained with explicit cues learned more than those who saw the nouns with the same orthographic cue highlighted with capital letters.  This effect persisted a week after training.  Testing with time pressure had the expected interfering effect on d-prime, but contrary to our predictions, explicit instruction was not more vulnerable to disruption in time pressure contexts (either training or testing under time pressure). 
  
==Explanation==
+
This could indicate that the learners in this study were not forced to use procedural knowledge with a 1400 ms deadline, and that a shorter time limit would show the expected problems with explicit instruction. However, it could also mean that instructional explicitness does not lead to representations that are less flexible.  By replicating the advantage explicit instruction had over correctness feedback while adding a non-explicit highlighting to the same cue information that was being explicitly taught, we showed that simple explicit cues can improve performance categorizing nouns by grammatical gender.
  
PENDING
 
 
==Descendents==
 
==Descendents==
  

Latest revision as of 21:29, 18 August 2010

PIs Presson, MacWhinney
Faculty MacWhinney
Postdocs Pavlik
Others with > 160 hours n/a
Study Start Date 10/01/08
Study End Date 05/30/08
Learnlab N/A
Number of participants (total) 80-100
Number of participants (treatment) N/A
Total Participant Hours ~100
Datashop? Summer '09

Abstract

This study extends the research question in a previous study by Presson & MacWhinney. In that experiment, explicit instruction led to better generalization and robustness to forgetting, compared to correct/incorrect feedback only. In the current study, explicitness is manipulated as well, with the more explicit condition the same as the previous study. However, the more implicit condition in this case is a highlighting manipulation, with no rules presented and in which participants see the relevant cue in capital letters to draw attention to the endings of words (e.g. "fromAGE" v. "fromage"). We want to know first if students are as effective at extracting cue patterns with this intervention as with strictly explicit instruction.

Second, we manipulate the time pressure of the task between groups. Time pressure favors procedural and automatic performance. If the highlighting condition is less explicit and equally effective, we expect an advantage within the time pressure condition (answer within 1400ms) compared to the no time pressure condition (trial times out at 6000ms, as in prior studies). This is important because time pressure is an essential element of naturalistic language use. In addition, time pressure changes the task demands, which we hypothesize will encourage more implicit representations of categories than with no time pressure.

Glossary


Research question

This research is designed to discover the best method of producing robust learning of French nominal gender, as well as the factors that make this learning more difficult.

Background and significance

Tucker, Lambert and Rigault (1977) evaluated the L1 (first language) learning of cues to gender in French. More recently, Holmes and Dejean de la Batie (1999) produced the first study of the acquisition of grammatical gender by L2 learners. Holmes and Segui (2004) have extended the detail of these analyses, but so far only with native speakers. Carroll (1999) and Lyster (2006) have explored the role of cue validity and availability in predicting usage by learners. All of these studies underscore the importance of high validity cues for the general vocabulary. However, these cues are only marginally useful for the highest frequency forms, whose gender must be learned more or less by rote. These analyses are in very close accord with the claims of the Competition Model (MacWhinney 1978, 2006).

In the Competition Model, each cue has a strength that is based on its reliability in signaling information (as in, for example, the use of spelling to predict grammatical gender). Some cues are more reliable than others: for instance, in the case of nouns that refer to people, semantic cues (the gender of a person) are more reliable than spelling cues. Over time, a learner picks up on these reliabilities, first acquiring the most clearly reliable cues, then later pulling apart conflicting but frequently co-occurring ones. Cue conflicts are then resolved through a process of competition. A full discussion of cue conflict is found in MacDonald and MacWhinney (1991).


In the current study, we manipulate whether participants receive explicit cues to gender patterns, or merely correct / incorrect feedback with the relevant cues highlighted (through capitalization) to direct attention.

Dependent variables

Accuracy and latency in training, as well as pre-/post-test gain scores, are dependent measures. In the post-test, there are two blocks: visual presentation (as in training) and auditory presentation (a transfer task).

Independent variables

Between-groups manipulations of time pressure (present or absent 1400ms time limit) and explicitness produce four groups:

Rule + Time Pressure

Rule - Time Pressure

Highlighting + Time Pressure

Highlighting - Time Pressure

Time pressure is operationalized by timimg out test trials after (some interval) ms.


Screenshots

Hypotheses

After 40 minutes of training, we hypothesize learning that is retained after a one-week retention interval and which generalizes to novel words under the same cue categories.

From previous studies, the response form has changed from seeing nouns in isolation and categorizing them as masculine or feminine to seeing both the correct and incorrect noun phrases on each trial, including a gendered definite article (e.g. "le fromage" and *"la fromage"). This method change may make learning more difficult than prior studies; however, it improves external validity by making the categorization process better resemble language input.

Because prior studies (e.g., Explicitness and Category Breadth) suggest that explicitness of instruction does not affect accuracy in training, we predict equivalent performance in the learning block. Explicit presentation of the prompt would be more robust to decay and generalization if in fact the manipulation makes the cue more salient (compared to the highlighting condition).

Adding time pressure will suppress initial performance in training. We predict that this suppression will be temporary, and in fact that later performance could show a time pressure advantage, especially in the no-feedback post-test. In addition, time pressure should change the relative effectiveness of an explicit inference rule vs. noticing highlights, such that noticing is easier than inference under time pressure, therefore improving learning.

At the one-week delay, participants complete an auditory version of the categorization task, testing whether the cue information transfers at all to correct French phonological representation, is too distant from French to help with auditory categorization, or is purely orthographic / visual.

Findings

Analyses were performed using d-prime as a measure of categorization sensitivity that accounts for possible response bias, and using response latency for correct responses.

  1. Pre-test performance was close to chance for all groups (mean accuracy = .51, SD= .11; highest group mean d-prime = 0.30, SD= 0.19).
  2. All groups learned the cues to grammatical gender as demonstrated by improved post-test d-prime and reduced post-test latency, both of which remained at a one-week delayed post-test.
  3. Explicit instruction showed higher accuracy than a less explicit highlighting condition, on tests both with time pressure (d-prime of explicit = 2.04, highlighting = 1.45) and without (d-prime of explicit = 2.70, highlighting = 1.96). This contradicts our speculation that time pressure might mitigate the positive effect
  4. Both explicit and highlighting instructional groups had lower d-prime when tested under time pressure.
  5. Training under time pressure did not eliminate the impairment in accuracy from testing under time pressure (i.e., both groups had higher d-prime without time pressure).
  6. Training under time pressure did not affect d-prime at any post-test.
  7. After controlling for mean pre-test latency, neither instructional variable (explicitness or training time pressure) influenced the amount of speed-up from pre-test to post-test. However, there was a marginally significant interaction (F(1,76) = 3.16, p = .08) such that the time pressure training responded faster at immediate post-test and that advantage disappeared after a one-week delay.

Explanation

Participants in all groups learned to categorize nouns by grammatical gender based on spelling cues. However, those who trained with explicit cues learned more than those who saw the nouns with the same orthographic cue highlighted with capital letters. This effect persisted a week after training. Testing with time pressure had the expected interfering effect on d-prime, but contrary to our predictions, explicit instruction was not more vulnerable to disruption in time pressure contexts (either training or testing under time pressure).

This could indicate that the learners in this study were not forced to use procedural knowledge with a 1400 ms deadline, and that a shorter time limit would show the expected problems with explicit instruction. However, it could also mean that instructional explicitness does not lead to representations that are less flexible. By replicating the advantage explicit instruction had over correctness feedback while adding a non-explicit highlighting to the same cue information that was being explicitly taught, we showed that simple explicit cues can improve performance categorizing nouns by grammatical gender.

Descendents

Annotated bibliography

  • Anderson, J. R., & Fincham, J. M. (1994). Acquisition of procedural skills from examples. Journal of Experimental Psychology: Learning, Memory, & Cognition, 20(6), 1322-1340.
  • Carroll, S. (1999). Input and SLA: Adults' sensitivity to different sorts of cues to French gender. Language Learning, 49, 37-92.
  • DeKeyser, R. M. (2005). What Makes Learning Second-Language Grammar Difficult? A Review of Issues. Language Learning, 55(Suppl1), 1-25.
  • Holmes, V. M., & Dejean de la Batie, B. (1999). Assignment of grammatical gender by native speakers and foreign learners of French. Applied Psycholinguistics, 20, 479-506.
  • Holmes, V. M., & Segui, J. (2004). Sublexical and lexical influences on gender assignment in French. Journal of Psycholinguistic Research, 33, 425-457.
  • Lyster, R. (2006). Predictability in French gender attribution: A corpus analysis. French Language Studies, 16, 69-92.
  • MacDonald, J. L., & MacWhinney, B. (1991). Levels of learning: A microdevelopmental study of concept formation. Journal of Memory and Language, 30, 407-430.
  • MacWhinney, B. (2006). A unified model. In N. Ellis & P. Robinson (Eds.), Handbook of Cognitive Linguistics and Second Language Acquisition. Mahwah, NJ: Lawrence Erlbaum Press.
  • Pimsleur, P. (1967). A memory schedule. The Modern Language Journal, 51(2), 73-75.
  • Robinson, P. (1997). Generalizability and automaticity of second language learning under implicit, incidental, enhanced, and instructed conditions. Studies in Second Language Aquisition, 19(2), 223-247.
  • Segalowitz, S., Segalowitz, N., & Wood, A. (1998). Assessing the development of automaticity in second language word recognition. Applied Psycholinguistics, 19, 53-67.
  • Ullman, M. T. (2001). The neural basis of lexicon and grammar in first and second language: the declarative/procedural model. Bilingualism: Language and Cognition, 4(1), 105-122.