Procuring digital preservation CAN be quick and painless with our new dynamic...
Second Language Anxiety and Achievement: A Meta-analysis
1. Second Language Anxiety and
Achievement: A Meta-analysis
YASSER TEIMOURI
GEORGETOWN UNIVERSITY
JULIA GOETZE
WASHINGTON UNIVERSITY
IN ST LOUIS
LUKE PLONSKY
NORTHERN ARIZONA UNIVERSITY
1
2. Introduction
A negative correlate of linguistic and non-linguistic variables:
Cognitive processing
Motivation
L2 Willingness to Communicate (L2 WTC)
Language achievement
Second Language Anxiety
Most researched affective variable in SLA
Aida, 1994; Hashimoto, 2002; Horwitz, 2010, 2017; Khajavy, MacIntyre, Barabadi, 2018; MacIntyre, 2017; MacIntyre & Gardner,
1994a, 1994b; Teimouri, 2017; Young, 1986
2
3. Introduction
Scovel (1978)
A turning-point
Noted inconsistent findings
for L2 anxiety
Methodological and theoretical issues
Measurement tools for L2 anxiety
MacIntyre & Gardner (1991)
Horwitz (2001)
Links between L2 anxiety and achievement
Potential sources of anxiety
Review of L2 anxiety research:
I. confounding phase
II. specialized phase
III. dynamic phase
MacIntyre (2017)
3
4. Introduction
The previous reviews can be extended in four areas:
Lack of systematicity and comprehensiveness (i.e., non-exhaustive selection of studies)
Emphasis on use of various anxiety measures (e.g. FLCAS, FLRAS) but not on language
achievement measures (e.g., course grades, language tests, GPA)
Only general discussion of the strength of the relationship between L2 anxiety
and achievement
No systematic investigation of moderators (e.g., context, educational level, target language)
4
5. Objective
Meta-analyze the relationship between L2 anxiety and language achievement:
To consolidate findings of previous L2 anxiety research
To address new research questions from a meta-analytic
perspective
5
6. Research questions
What are the methodological and reporting practices
involved in studies of the relationship between L2 anxiety
and achievement?
What is the direction and the size of the relationship
between anxiety and L2 achievement?
How does this relationship vary across different language
achievement measures, learning contexts, L2 anxiety types,
and educational levels?
6
8. Literature search
Education Resources Information
Clearinghouse (ERIC),
anxiety, language, second (L2),
foreign (FL), classroom,
communication, use, emotion,
affect, and achievement Google Scholar
Linguistics and Language
Behavior Abstracts (LLBA),
ProQuest,
Manual SearchReferences Forward citation
700+
documents
Literature Search
Published and unpublished studies
8
9. Inclusion/Exclusion criteria
Eligibility criteria
A measure of language anxiety and a measure of L2
achievement;
8
Dissertations
89
Journal articles
Correlation between language anxiety and
language achievement;
Published after Horwitz. (1986)
9
K = 97
13. The database
19,933 students
(M = 190; SD = 222)
A total of 105 independent samples
(Sample sizes: 12 to 1,174)
204 effect sizes (r)
97 studies
(1985-2017)
23 countries
13
14. Anxiety measures
Reliability estimates for measures of L2 anxiety
Instruments K Reported Missing Min Max Mean SD
All L2 anxiety measures 129 84 45 .65 .97 .88 .06
129 anxiety questionnaires
25 different anxiety scales
14
RQ1
15. Anxiety measures
Reliability estimates for the three most frequently used measures of L2 anxiety
Instruments K Reported Missing Min Max Mean SD
Horwitz et al. (1986) 67 40 27 .75 .97 .90 .05
Saito et al. (1999) 11 8 3 .82 .96 .87 .04
Gardner (1985) 9 6 3 .86 .92 .88 .02
15
RQ1
129 anxiety questionnaires
25 different anxiety scales
16. Language achievement measures
90 language tests 5 GPAs40 course grades 35 self-perceived
25.3% reported Mean (k = 43)
22.9% reported SD (k = 39)
40.6% range of possible scores (k = 69)
170 language measuresLanguage
16
RQ1
17. Language achievement measures
Measures K Reported Missing Median Mean SD
Language tests 90 10 80 .90 .85 .13
Self-perceived
proficiency
35 13 22 .89 .87 .06
Total 125 23 102 .89 .86 .10
Reliability analyses of language tests and self-assessment measures
17
RQ1
18. Correlation between anxiety and achievement
95% CIs
K N Mr Lower Upper
L2 achievement* 96 19,933 -.36 -.39 -.33
Note. p < .001
*Excluding 9 outliers (standard residuals > 2.5)
Language
18
RQ2
19. Moderator analysis: Achievement measures
95% CIs
Achievement measure K N Mr Lower Upper
Self-perceived 8 3360 -.47 -.54 -.40
Course grades 30 7129 -.34 -.39 -.29
Language test 44 10388 -.36 -.39 -.32
GPAs 5 878 -.26 -.50 -.02
19
RQ3
20. 95% CIs
Study Context K N Mr Lower Upper
Study abroad 2 214 -.49 -.66 -.37
Second language 21 2846 -.40 -.46 -.33
Foreign language 72 15973 -.35 -.39 -.31
20
RQ3
Moderator analysis: Study context
21. 95% CIs
Educational Level K N Mr Lower Upper
Elementary 3 832 -.43 -.79 -.06
Junior high school 8 2053 -.34 -.49 -.19
Senior high school 27 5508 -.34 -.41 -.27
College 53 9811 -.37 -.42 -.32
Language institute 2 518 -.25 -.65 .15
21
RQ3
Moderator analysis: Educational levels
22. 95% CIs
Target language K N Mr Lower Upper
English 62 15666 -.33 -.38 -.29
Non-English 34 3392 -.39 -.46 -.33
22
RQ3
Moderator analysis: Target language
25. RQ1: Methodological and reporting practices
Anxiety measures
Language-specific definition
and operationalization
Fairly reliable instruments
Achievement measures
Lower levels of rigor and
transparency
Language
Reliability ≠ validity
• Sparks & Patton (2013)
• Teimouri (2018)
More attention needed
• Brown, Plonsky, & Teimouri (2018)
• Thomas (2006)
• See Ziegler et al. (in prep)
25
26. RQ2: Anxiety & L2 achievement?
Anxiety
r = −.36
L2 achievement
13% of variance
Medium/moderate effect
Plonsky & Oswald (2014)
r = −.21
Seipp (1991)
Education Mathematics
r = −.27
Ma (1999)
26
27. 0.49
0.37
-0.36
0.31
0.26
0 0.1 0.2 0.3 0.4 0.5 0.6
Aptitude
Motivation
Anxiety
Strategies
WM
Meta-analytic correlations (r)
RQ2: Anxiety & L2 achievement?
Li (2016)
M&G (2003)
THIS STUDY
Plonsky (2011)
Linck et al. (2014)
27
28. RQ3: Moderator analysis: Achievement measures
MacIntyre et al. (1997)
Anxious students tend to underestimate their language competence
Self-assessment: Strongest relationship
Language tests and course grades: Anxiety has similar negative effects
Brown et al. (2018)
Both good measures to capture negative influence of L2 anxiety
GPAs: Smallest relationship
GPA NOT a language
achievement measure
28
29. Each context has its own unique,
inherent challenges
(SL vs. FL) context
Educational Levels
Two confounding variables:
- Age of students
- Unique features of each context
No difference Mixed results!
29
RQ3: Moderator analysis: Context and educational levels
30. RQ3: Moderator analysis: The target language
Are English learners just special?
Less negative effects for English learners
(vs. non-English learners)
Unique features of the target
language (Chinese vs. English)
+ Familiarity with English
(English as an international language)
30
31. RQ3: Moderator analyses: Anxiety types
Strongest negative effects on
listening performance
No time pressure;
More resources (e.g., dictionary)
Lowest negative effects on
reading performance
Time pressure!
Difficult to measure!
31
32. Limitations
Small sample sizes
(e.g., study abroad context,
elementary students)
Exclusion of non-English
literature
Sequencing of languages
(L2– L5)
32
33. Future research
Apply more comprehensive statistical analyses, such as multiple
regression and structural equation modeling (see Plonsky & Ghanbar, in press)
Correct attenuation effects of scales with low
reliability (see Plonsky & Derrick, 2016; Teimouri, 2017).
Use a combination of achievement measures
Caution about the use of GPAs
Improve reporting of statistical analyses
Sub-optimal measures of L2 achievement might
attribute to inconsistent and attenuated results
33
Dynamic (real-time) + physiological measures of anxiety (and performance)
34. We can put to bed, once and for all, whether or not
anxiety is best conceptualized as a negative predictor of
language achievement (MacIntyre, 2017)
Focus on novel theoretical perspectives
and robust measurement tools
Conclusion
34
35. Teimouri, Y., Goetze, J., & Plonsky, L. (in press). Second language anxiety and achievement: A meta-analysis.
Studies in Second Language Acquisition.
35
36. JULIA GOETZE
WASHINGTON UNIVERSITY IN
ST LOUIS
JULIA.GOETZE@WUSTL.EDU
LUKE PLONSKY
NORTHERN ARIZONA UNIVERSITY
LUKEPLONSKY@GMAIL.COM
YASSER TEIMOURI
GEORGETOWN UNIVERSITY
YT329@GEORGETOWN.EDU
36
Thank you!
38. Publication Bias
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
0.0
0.1
0.2
0.3
0.4
StandardError
Fisher's Z
Funnel Plot of Standard Error by Fisher's Z
trim-and-fill procedure
Duval and Tweedie (2000)
Right side: r = – .30Left side: r = – .36
38
39. RQ2: What is the direction and the size of the relationship between anxiety
and learners’ language achievement?
r =.37
Masgoret & Gardner (2003)
L2 Motivation
r = .47
Li (2016)
Language Aptitude
39
Anxiety
r = −.36
L2 achievement
13% of variance
Medium/moderate effect
Plonsky & Oswald (2014)