The Efficacy of Video Feedback on SelfEvaluation of Performance and Treatment of
Bilingual Participants: A Linguistically and
Culturally Sensitive Intervention for Public
Speaking Anxiety
Magdalena Perez
July 25, 2006
Introduction (Cont.)
Purpose of Study
 The primary objective of the current study
was to examine the combined impact that
treatment-type (bilingual-video feedback
vs. monolingual- video feedback) and
audience-race/ethnicity (Latino vs. White)
had on the treatment outcome of bilingual
speech anxious participants.
Introduction (Cont.)
Studies Supporting the Video Feedback
Intervention:

“Pure” Video Feedback:

Speech performances are rated more positively
(Rapee & Hayman, 1996).
Cognitive Preparation + Video Feedback:


Speech performances are rated



more positively (Harvey et al., 2002; Kim,et al. 2002)
more accurately (Harvey et al., 2002)
less anxiety provoking (Kim et al., 2003; Smits, 2004)
Introduction (Cont.)

Why should we even consider applying video
feedback in a bilingual fashion?


Bilinguals experience different levels of anxiety
according to the language which they are
speaking (Stein et al., 1998)
Why should we even consider manipulating
the race/ethnicity of audience members?

Minority bilinguals are susceptible to stereotype
threat (Stein et al., 1998).
Hypotheses
Language Specific:

H1: Participants were expected to
experience differential levels of
anxiety when performing in
English versus Spanish at pretreatment.

H2: Participants were expected to
evaluate their English behavioral
speech performance differently
than their Spanish behavioral
speech performance at pretreatment.
Within-Group Difference
Video Feedback:



H3: Participants were expected to
experience a greater reduction on
a state measure of public
speaking anxiety at posttreatment and follow-up.
H4: Participants were expected to
rate their behavioral speech
performances more positively at
post-treatment and follow-up.
H5: Participants were expected to
rate their behavioral speech
performances more accurately at
post-treatment and follow-up.
Exploratory Questions on BetweenGroup Differences
Public Speaking Measures

EQ1: The present study would
attempt to identify the
combination of treatment-type
and audience-race/ethnicity
(Treatment x Ethnicity) that
would show the greatest
degree of improvement on
public speaking outcome
measures.
Social Anxiety Measures

EQ 2: The present study
would attempt to identify the
combination of treatment-type
and audience-race/ethnicity
(Treatment x Ethnicity) that
would show the greatest
degree of improvement on
social anxiety outcome
measures.
Exploratory Questions on BetweenGroup Differences (Cont.)
Observers’ Ratings
 EQ3: The present study would
attempt to identify the
combination of treatment-type
and audience-race/ethnicity
(Treatment x Ethnicity) that
showed the greatest degree of
improvement based on
observers’ ratings of
participants’ speech
performance and anxiety.
Observer—Participant
Discrepancy Scores:
 EQ4: The present study would
attempt to determine the
combination of treatment-type
and audience-race/ethnicity
(Treatment x Ethnicity) that
had participants rate their
speech performance and
anxiety scores more similarly
to those ratings of
independent observers.
Methods
Participants

Criteria





Mexican descent
Bilingual
≥9 LSAS-SR public speaking fear subscale
≥ 50 SUDS ratings for each baseline speech
Recruitment

301 Psychology subject pool & pre-screening
website
Methods (Cont.)
1st
Session
Bilingual Tx-Anglo Bilingual Tx-Latino English Tx-Anglo
English Tx-Latino
PT: Eng
PT : Span
PT: Eng
PT: Span
PT: Eng
PT: Span
PT: Eng
PT: Span
4 tx speeches (E)
4 tx speeches (E)
4 tx speeches (E)
4 tx speeches (E)
--------------------------------------------------------------------------------------------------------2nd
Session
4 tx speeches (S)
4 tx speeches (S)
4 tx speeches (E)
4 tx speeches (E)
PA: Eng
PA: Eng
PA: Eng
PA: Eng
PA: Span
PA: Span
PA: Span
PA: Span
--------------------------------------------------------------------------------------------------------3rd
Session
FU: Eng
FU: Span
FU: Eng
FU: Span
FU: Eng
FU: Span
FU: Eng
FU: Span
Methods (Cont.)
Measurements
Public Speaking State
Measures:



Appraisal of Social Concerns Scale
(ASC; Lucas, 1993)
Subjective Units of Distress (SUDS)
Speech Evaluation Questionnaire
(SEQ; Harvey, et al., 2000)
Public Speaking Trait
Measures:


Liebowitz Social Anxiety SelfReport Scale-Public Speaking
Subscale (LSAS-PS; Watson &
Friend, 1969)
Speech Anxiety Thoughts
Inventory (SATI; Cho, Smits, &
Telch, 2004)
Methods (Cont.)
Measurements
Social Anxiety Constructs:

Liebowitz Social Anxiety ScaleSelf Report (LSAS-SR;
Liewbowitz, 1987)

Fear of Negative Evaluation
(FNE; Watson & Friend, 1969)
General Psychological Measure:
 Brief Symptoms Inventory (BSI;
Derogatis, 1993)
Methods (Cont.)
Measurements
Treatment Expectancy:
 Reaction to Treatment
Questionnaire (RTQ; Borkovec
& Nau, 1972)
Cultural Characteristics:
 Woodcock-Munoz Language
Survey (WMLS; Woodcock &
Munoz, 2001)
 Acculturation Rating Scale for
Mexican Americans-II (ARSMAII; Cuellar, Arnold, &
Maldonaldo, 1995)
 Family Attitude Scale-Revised
(FAS-R; Carrasco, 1990)
 Multidemensional Acculturaive
Stress Inventory (MASI;
Rodriguez et al., 2002)
Methods
Assessor Ratings
 Two assessors


Blind to the purpose of the experimental
design of the study
Rated all the pre-treatment, post-treatment,
and follow-up speech performance using the
SEQ & SUDS.
Statistical Analyses

Baseline differences:

Language Specific Outcome Variables


Non-language Specific Outcome Variables


2 (Treatment) x 2 (Audience) ANOVA
Covariates:


2 (Treatment) x 2 (Audience) x 2 (Language) ANOVA
If correlation ≥ .70 between change scores and the variable
being examined as a potential covariate then it was decided
to that this variable would be entered in as a covariate.
General Treatment Outcome:

All outcome analyses were intent-to-treat where the last
observation was carried forward.


Within-group: Repeated Measures MANOVA
Between-group: Depending on the measure:
 2 (Treatment) x 2 (Audience) x 2 (Language) ANCOVA
 2 (Treatment) x 2 (Audience) ANCOVA
Results
Sample Characteristics

Total of 46 participants were randomly assigned to one of the
conditions








Mean age = 21.9
Cultural Identity: Slightly oriented Anglo American bicultural identity
Family Values: Bicultural family values with a slight orientation
toward White European values
Language Proficiency:



Bilingual Video Feedback White Audience: 9
Bilingual Video Feedback Latino Audience: 14
English Video Feedback White Audience: 10
English Video Feedback Latino Audience: 13
English: Average range
Spanish: Borderline range
Socioeconomic Status:$35,000 to $45,000 parent’s household
annual income
Result (Cont.)

Drop outs:


Before end of treatment: 1 EVW; 3 EVL; 3 BVW; 3 BVL
After post-assessment: 1 BVW & 1 BVL





Drop outs rates did not vary across the conditions.
No differences on the baseline measures between the completers
and non-completers
Four conditions did not vary on any of the outcome
measures at pre-treatment.
UT and SEU did not vary on any of the outcome
measures at pre-treatment.
No covariates were identified
Results (Cont.)
H1: Differential Levels of
Anxiety
 Paired sample t-test: ESUDS vs. S-SUDS

Participants on average
reported higher SUDS
levels when giving their
pre-treatment speech in
Spanish (M = 77.17, SE =
2.41) than English (M =
71.17, SE = 2.38), t (45) =
-2.34, p < .05.
H2: Differential Levels of
Performance Ratings:
 Paired sample t-test: ESEQ vs. S-SEQ

Participants rated their
baseline Spanish speech
performance more poorly
(M = 90.67, SE = 5.51)
than their English speech
performance (M = 101.96,
SE = 6.13), t (45) = 2.20,
p < .05.
Results (Cont.)
H3: W/in Grp Diff. on Video
Feedback on SUDS
 MANOVA

Participants in all the
conditions experienced a
significant decline at posttreatment and follow-up on
their English-SUDS and
Spanish-SUDS scores (ps <
.05).
H4: W/in Grp Diff. on Video
Feedback on SEQ
 MANOVA

Participants in all the
conditions rated their
English and Spanish
speeches more positively at
post-treatment and followup than they had at pretreatment (ps < .05).
Results (Cont.)
H5: W/in Grp Diff on SEQ Accuracy Ratings
 MANOVA

Participants in all the conditions rated their English speech performance
more accurately at post-treatment and follow-up (ps < .05).

Only participants in the BVL and EVL conditions rated their Spanish
speeches more accurately at post-treatment and follow-up (ps < .05).

Participants in the EVW condition only rated their Spanish speech more
accurately at follow-up (p < .05), but not at post-treatment.

Participants in the BVW condition did not rate their Spanish speeches
more accurately at post-treatment or follow-up (ps > .05).
Results (Cont.)
EQ1: Public Speaking State Outcome Measures:
Pre to Post-treatment

Treatment x Audience X
Language ANCOVA



Concerns About Giving
Speech (ASC): (F (1, 40) =
1.28, p > .05)
Peak Anxiety Level (SUDS): (F
(1, 40) = 0.64, p > .05)
Self-Performance Ratings
(SEQ): (F (1,40) = 1.22, p
>.05)
Pre-treatment to Follow-up

Treatment x Audience x
Language ANCOVA



Concerns About Giving
Speech (ASC): (F (1, 40) =
1.87, p > .05)
Peak Anxiety Level (SUDS): (F
(1, 40) = 0.89, p > .05)
Self-Performance Ratings
(SEQ): (F (1, 40) = 2.71, p >
.05)
Results (Cont.)
EQ1: Public Speaking Trait Outcome Measure: Public Speaking Anxiety (LSAS-PS)
Pre-treatment to Follow-up
Pre to Post-treatment


Treatment x Audience: (F (1, 41)
= 5.92, p < .05)
25
Bilingual Treatment
English Treatment
21.56
21.01
21.67
LSAS-PS (0-30)
20
15
16.2
10
5
0
White Audience (EVW/BVW)
Latino Audience (EVL/BVL)
Treatment x Audience: F (1, 41) =
3.76, p > .05
Results (Cont.)
EQ1: Public Speaking Trait Outcome Measure: Speech Anxious Thoughts (SATI)
Pre to Post-treatment

Treatment x Audience x Language
ANCOVA: (F (1, 40) = 0.62, p >
.05)
 Audience X Language (F (1,
40) = 4.11, p < .05)
84
82.31
82
SATI-English
SATI-Spanish
80
SATI (0-115)
78
77.51
76
74
72
71.49
70
70.3
68
66
64
White Audience (EWA/BWA)
Latino Audience (EVL/BVL)
Pre-treatment to Follow-up

Treatment x Audience x
Language: F (1, 40) = 0.55. p >
.05
 Audience x Language: F (1,
40) = 3.02, p > .05
Results (Cont.)
EQ2: Social Anxiety Outcome Measures:Fear of Negative Evaluation(FNES)
Pre- to Post-treatment

Pre- to Follow-up
Treatment x Audience: F (1, 41) =
2.93, p > .05

Treatment x Audience: F (1, 41) =
3.43, p > .05
 Audience: F (1, 41) = 4.33, p
< .05
23

Audience: F (1, 41) = 4.42, p <
.05
23.5
FNE Scores
FNE Scores
22.34
22.5
22.96
23
22
21.5
22.5
21
FNE (0-30)
FNE (0-30)
22
21.5
21
20.5
20
19.5
19.43
20.78
19
20.5
18.5
20
18
17.5
19.5
White Audience (EVW/BVW)
Latino Audience (EVL/BVL)
White Audience (EVW/BVW)
Latino Audience (EVL/BVL)
Results (Cont.)
EQ2: Social Anxiety Outcome Measures: General Social Anxiety (LSAS-Total)


Pre- to Post-treatment
Pre-treatment to Follow-up
Treatment x Audience: F (1, 41) =
4.44, p < .05
Treatment x Audience: F (1, 41) =
5.15, p < .05

80
80
Bilingual Treatment
English Treatment
Bilingual Treatment
English Treatment
72.89
70
71.86
70
71.17
68.47
67.27
LSAS-Total (0-144)
LSAS-Total (0-144)
70.1
60
50
50.46
60
50
40
40
30
30
20
20
10
10
0
44.93
0
White Audience Conditions (BVW/EVW)
Latino Audience Conditions (BVL/EVL)
White Audience Conditions (BVW//EVW) Latino Audience Conditions (BVL/EVL)
Results (Cont.)
Preliminary Analyses with Observers Ratings



Only 72% of pre-treatment, post-treatment, and followup speeches were evaluated by the observers.
Reliability of observers ratings on performance (SEQ)
and anxiety ratings (SUDS) was good (ranged .71 to .81)
Baseline Difference Across Four Groups

Treatment x Audience x Language: non-significant for
performance and anxiety ratings

Audience x Language: significant for performance ratings

Participants in front of a White audience were rated more poorly on
English performance than those in front of a Latino audience.
Results (Cont.)
EQ3: Observer Ratings: Peak Anxiety Ratings (SUDS)


Pre- to Post-treatment
Treatment x Audience x
Language: F (1, 15) = 0.04, p >
.05
Pre-treatment to Follow-up

Treatment x Audience x
Language: F (1, 13) = 5.01, p <
.05

Insufficient power in model
Results (Cont.)
EQ3: Observer Ratings: Performance Ratings (SEQ)


Pre- to Post-treatment
Treatment x Audience x
Language: F (1, 15) = 0.35, p >
.05
Pre-treatment to Follow-up

Treatment x Audience x
Language: F (1, 13) = 6.5, p <
.05
 Insufficient power in model
Results (Cont.)
Preliminary Analyses with Observer-Participants
Discrepancy Scores

Peak anxiety (SUDS) and performance (SEQ)
deviation scores


Observers’ mean score — participant’s score
Baseline Difference Across Four Groups

Treatment x Audience x Language: non-significant for
anxiety and performance ratings
Results (Cont.)
EQ4: Observer-Participants Discrepancy Scores: Peak Anxiety Ratings (SUDS)
Pre- to Post-treatment

Treatment x Audience x
Language: F (1, 15) =.001, p >
.05
Pre-treatment to Follow-up

Treatment x Audience x
Language: F (1, 12) = 0.57, p >
.05
Results (Cont.)
EQ4: Observer-Participants Discrepancy Scores: Performance Ratings (SEQ)
Pre- to Post-treatment

Treatment x Audience x
Language: F (1, 15) = 1.37, p >
.05
Pre-treatment to Follow-up

Treatment x Audience x
Language: F (1, 12) = 4.77, p <
.05

Insufficient power in model
Post-hoc Analyses
Observers’ Ratings English vs. Spanish
 Observers did not differentially
rate participants’ English-SUDS
scores from their Spanish- SUDS
scores at pre-treatment.
 Observers rated their Spanish
speech performance more poorly
on the SEQ measure than their
English speech performance at
pre-treatment.
Accuracy of Participants Ratings
 Participants were equally
accurate at rating their
English and Spanish
speeches at pre-treatment
on peak anxiety (SUDS)
and performance (SEQ)
measures.
Post-hoc Analyses (Cont.)
Eight Exposure Speeches
 2 x 2 ANOVA:
 ASC (F (7, 36) = 1.08, p >
.05)
 SUDS (F (7, 36) = 1.84, p
> .05
 SEQ (F (7, 36) = 1.32, p
>.05)
Family Values & Cultural Identities
 Regressional analyses
indicated that the degree of
identification with Mexican
American family values and
cultural identity type were not
predictive of treatment
outcome.
 Regressional analyses
indicated that family values
and cultural identity
orientation were not related to
the degree to which
participants accurately
perceived their speeches.
Discussion


Participants reported language specific
anxieties and self-performance ratings at
pre-treatment.
Participants were equally accurate at
evaluating their English and Spanish
speeches at pre-treatment.

Their degree of accurately evaluating their
speeches is not related to their orientation on
family values or cultural identity.
Discussion (Cont.)


Consistent with previous findings (Kim et al., 2002; Smits, 2004;
Rapee & Hayman, 1996), participants in the present study
experienced a reduction on heightened anxiety
levels and tended to rate their speeches more
positively following treatment.
Contrary to previous findings (Harvey et al., 2000; Rapee & Hayman 1996)
not all participants were able to rate their
speeches more accurately over time following
the intervention.
Discussion (Cont.)

Bilingual treatment did not show any
advantageous gain above and beyond the
English-only treatment.


Lack of language-specific inaccurate
perceptions
Participants may not have had pre-determined
perceptions of their Spanish public speaking
abilities given their present cultural
environment.
Discussion (Cont.)

White audience conditions contributed
substantially to some of the findings.


Did participants in the White audience
conditions benefit from stereotype threat?
Did participants in the White audience
conditions benefit from how closely the
exposure task resembled their actual fear
structure?
Clinical Implications

Consistent with the recommendations purposed by the
APA’s Multicultural Psychotherapy Guidelines (APA,
2002), the findings of this study emphasize the
importance of focusing on the client within his or her
cultural context.


Cognitive and Cultural Flex Theory (Ramirez, 1999) states that
individual belonging to more than one culture have the potential
to “flex” culturally and cognitively to adjust to their cultural
environments.
The key is to identify if the participants are having difficulty
“flexing” to one of their cultural environments and structuring
the exposure accordingly (match vs. mismatch environment).
Clinical Implications (Cont.)

Findings also address another recommendation
made by APA’s Multicultural Psychotherapy
Guidelines (APA, 2002): identifying culturalspecific strategies for interventions


Do bilinguals have language specific inaccurate
perceptions?
Do bilinguals’ level of anxiety vary according to the
ethnic composition of the audience members?
Limitations





The present study was underpowered because of the
modest cell sizes.
Generalizability of findings is limited.
The range of family values and cultural identities was
restricted.
Language proficiency of sample was restricted.
The study failed to collect data on whether or not
participants were a) consciously aware of the
racial/ethnic composition of their audience members and
b) how this may have influenced their treatment.
Descargar

The Efficacy of Cognitive Preparation Plus Video Feedback