MSc. Vilma Susana Romero RomeroVilma Susana Romero Romero VI Congreso Bayesiano de Am erica Latina...
Transcript of MSc. Vilma Susana Romero RomeroVilma Susana Romero Romero VI Congreso Bayesiano de Am erica Latina...
tugrazLancaster University
Longitudinal Multidimensional Item Response
Modelling in Preschool Children’s Mental State
Understanding
MSc. Vilma Susana Romero Romero
VI Congreso Bayesiano de America Latina
Vilma Romero VI COBAL1 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
Conclusionsand FutureWork
tugrazLancaster University
Theory of Mind (ToM)
Definition
Ability to perceive our own mental states as well as from
others, such as beliefs, desires and intentions and know that
they differ from one person to another.
Main Features
Developed in the first years of life (4 years old).
Understand social environment and how to interact in it.
Mental state tasks to identify the acquisition of ToM.
Let’s take a look: Aim
Understanding of mental states in
children over the 3rd year of life through
MIRT and analyze each dimension under
the Bayesian Longitudinal approach.
Vilma Romero VI COBAL2 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
Conclusionsand FutureWork
tugrazLancaster University
Data Description
Participants
86 British children (Female = 41, Male = 45) from different
preschools and day nurseries located in Northern Lancashire.
Age: Between 30 and 33 months when recruited.
Measures
8 mental state tasks (13 questions three times in intervals of 4
months). A correct response scored ‘1’ and an incorrect
response scored ‘0’.
Standard Location Change
Deceptive Box
Pretence, Desire and Think
Narrative
Verbal and Non-Verbal (2 and 4 trials repectively)
Vilma Romero VI COBAL3 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
Conclusionsand FutureWork
tugrazLancaster University
Exploratory Analysis
Response Patterns
1.1.0
1.0.1
0.1.0
1.1.1
1.0.0
0.1.1
0.0.1
0.0.0
NVFBTR1
NVFBTR2
NVFBTR3
NVFBTR4
VFBQ1T1
VFBQ1T2
NARTESQ
DBOTHEQ
DBSELFQ
STANTES
PRETENSE
DESIRE
THINK
5 2 2 4 3 0 12 2 2 0 1 0 1
2 6 6 2 2 1 9 1 4 2 10 8 3
5 6 1 4 5 8 5 8 10 10 7 4 12
8 3 2 5 3 2 15 2 5 4 24 20 4
5 6 5 3 8 6 12 14 12 5 11 8 4
11 13 18 12 6 5 10 4 6 9 8 13 12
17 16 18 15 14 18 12 8 8 13 11 13 15
13 14 14 21 25 26 9 45 37 41 12 18 33
0
5
10
15
20
25
30
35
40
45
Only complete observations
taken into account.
Most difficult tasks: Standard
Location Change and Deceptive
Box.
Total Performance Trend
0.00
0.25
0.50
0.75
1.00
1 2 3
Time
Suc
cess
Inde
x
General trend is increasing
Vilma Romero VI COBAL4 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
Conclusionsand FutureWork
tugrazLancaster University
Correlation Analysis
NVFBTR1
NVFBTR2
NVFBTR3
NVFBTR4
PRETENSE
DESIRE
THINK
NARTESQ
VFBQ1T1
VFBQ1T2
DBOTHEQ
DBSELFQ
STANTES
NVFBTR1
NVFBTR2
NVFBTR3
NVFBTR4
PRETENSE
DESIRETHINK
NARTESQ
VFBQ1T1
VFBQ1T2
DBOTHEQ
DBSELFQ
STANTES
1 0.2 0.2 0.4 −0.1 0 0.3 0.1 0.2 0.1 −0.1 0 0.1
0.2 1 −0.1 0.5 −0.4 −0.1 0.1 −0.4 0 0.2 0.1 0 0
0.2 −0.1 1 0.6 −0.1 0.1 0 0 −0.2 0.2 0.2 0.3 0.3
0.4 0.5 0.6 1 −0.2 0.3 0.1 −0.1 −0.1 0.3 0.3 0.2 0.2
−0.1 −0.4 −0.1 −0.2 1 0.6 0.2 0.2 0.4 −0.3 0.2 0.4 0.2
0 −0.1 0.1 0.3 0.6 1 0.1 0.2 0.3 0.1 0.1 0.3 0.3
0.3 0.1 0 0.1 0.2 0.1 1 0.3 0.4 0.3 0.1 0.1 0.4
0.1 −0.4 0 −0.1 0.2 0.2 0.3 1 0.3 −0.4 0 0.1 0
0.2 0 −0.2 −0.1 0.4 0.3 0.4 0.3 1 −0.1 0.1 −0.1 0
0.1 0.2 0.2 0.3 −0.3 0.1 0.3 −0.4 −0.1 1 0.1 −0.2 0.5
−0.1 0.1 0.2 0.3 0.2 0.1 0.1 0 0.1 0.1 1 0.4 0.4
0 0 0.3 0.2 0.4 0.3 0.1 0.1 −0.1 −0.2 0.4 1 0.3
0.1 0 0.3 0.2 0.2 0.3 0.4 0 0 0.5 0.4 0.3 1
−1.0
−0.5
0.0
0.5
1.0
NVFBTR1
NVFBTR2
NVFBTR3
NVFBTR4
PRETENSE
DESIRE
THINK
NARTESQ
VFBQ1T1
VFBQ1T2
DBOTHEQ
DBSELFQ
STANTES
NVFBTR1
NVFBTR2
NVFBTR3
NVFBTR4
PRETENSE
DESIRETHINK
NARTESQ
VFBQ1T1
VFBQ1T2
DBOTHEQ
DBSELFQ
STANTES
1 0.5 0.5 0.5 −0.1 0 0 −0.1 −0.1 −0.1 0 0 0
0.5 1 0.3 0.5 0.2 0.2 0.3 −0.2 0 0.2 0.1 0.2 0.4
0.5 0.3 1 0.6 0.3 0.2 0.1 0 0.4 −0.4 0.4 −0.1 0.3
0.5 0.5 0.6 1 0.2 0.4 0.5 0 0.1 −0.3 0.4 −0.1 0.3
−0.1 0.2 0.3 0.2 1 0.6 0.6 0.1 0.3 −0.4 0.1 0.1 0.4
0 0.2 0.2 0.4 0.6 1 0.6 0 0.2 0 0.3 0.2 0.4
0 0.3 0.1 0.5 0.6 0.6 1 0.3 0.1 0.1 0.5 0.2 0.1
−0.1 −0.2 0 0 0.1 0 0.3 1 0.1 0.2 −0.3 0 0.1
−0.1 0 0.4 0.1 0.3 0.2 0.1 0.1 1 0.6 0 −0.1 0.4
−0.1 0.2 −0.4 −0.3 −0.4 0 0.1 0.2 0.6 1 0 0.1 0.1
0 0.1 0.4 0.4 0.1 0.3 0.5 −0.3 0 0 1 0.7 0.3
0 0.2 −0.1 −0.1 0.1 0.2 0.2 0 −0.1 0.1 0.7 1 0.2
0 0.4 0.3 0.3 0.4 0.4 0.1 0.1 0.4 0.1 0.3 0.2 1
−1.0
−0.5
0.0
0.5
1.0
NVFBTR1
NVFBTR2
NVFBTR3
NVFBTR4
PRETENSE
DESIRE
THINK
NARTESQ
VFBQ1T1
VFBQ1T2
DBOTHEQ
DBSELFQ
STANTES
NVFBTR1
NVFBTR2
NVFBTR3
NVFBTR4
PRETENSE
DESIRETHINK
NARTESQ
VFBQ1T1
VFBQ1T2
DBOTHEQ
DBSELFQ
STANTES
1 0.8 0.8 0.7 0.5 0.5 0.3 0.4 0.2 0.4 0.4 0.4 0.2
0.8 1 0.7 0.8 0.6 0.5 0.4 0.2 0.2 0.6 0.2 0.4 0.2
0.8 0.7 1 0.7 0.4 0.3 0 0.4 0.2 0.3 0.1 0 0.3
0.7 0.8 0.7 1 0.5 0.4 0.2 0.2 0.1 0.3 0.2 0.3 0.4
0.5 0.6 0.4 0.5 1 0.9 0.7 0.6 0 0.2 0.3 0.6 0.6
0.5 0.5 0.3 0.4 0.9 1 0.8 0.3 0.3 0.3 0.4 0.4 0.4
0.3 0.4 0 0.2 0.7 0.8 1 0.3 0.3 0.3 0.3 0.5 0.5
0.4 0.2 0.4 0.2 0.6 0.3 0.3 1 0.2 0.5 0.2 0.2 0.4
0.2 0.2 0.2 0.1 0 0.3 0.3 0.2 1 0.7 0.2 0.3 0.4
0.4 0.6 0.3 0.3 0.2 0.3 0.3 0.5 0.7 1 0.4 0.1 0.3
0.4 0.2 0.1 0.2 0.3 0.4 0.3 0.2 0.2 0.4 1 0.6 0.2
0.4 0.4 0 0.3 0.6 0.4 0.5 0.2 0.3 0.1 0.6 1 0.5
0.2 0.2 0.3 0.4 0.6 0.4 0.5 0.4 0.4 0.3 0.2 0.5 1
−1.0
−0.5
0.0
0.5
1.0
Time 1 and 2 (Above)
Time 3 (Below)
Correlation across time.
Correlation between
possible latent factors.
Vilma Romero VI COBAL5 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
MIRT
Causality Analysis
Conclusionsand FutureWork
tugrazLancaster University
Multidimensional Item Response Modeling
The probability of answering a dichotomous itemcorrectly is:
Φ(xij = 1|θi , αj , dj) =1
1 + exp[−D(αTj θi + dj)]
Where, i = 1, ...,N participants, j = 1, ..., n test items,m latent factors θi = (θi1, ..., θim) with associated itemslopes αj = (α1, ..., αm), dj is the item intercept and Dis a scaling adjustment (usually 1.702).
The slopes are the multidimensional discriminationparameters (one for each latent factor).
The intercept is proportional to the item difficulty.
The higher (lower) the discrimination parameter, the (worst)better the item distinguishes low and high ability levels.
Vilma Romero VI COBAL6 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
MIRT
Causality Analysis
Conclusionsand FutureWork
tugrazLancaster University
MIRT as Item Factor AnalysisNVFBTR1
NVFBTR2
NVFBTR3
NVFBTR4
PRETENSE
DESIRE
THINK
NARTESQ
VFBQ1T1
VFBQ1T2
DBOTHEQ
DBSELFQ
STANTES
F1*
0.740.590.610.66
F2*0.63
0.61
0.36
F3*0.560.65
F4*0.88
0.47
g
0.31
0.42
0.43
0.55
0.53
0.65
0.57
0.28
0.41
0.37
0.37
0.39
0.71
Bifactor Model Path
Vilma Romero VI COBAL7 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
MIRT
Causality Analysis
Conclusionsand FutureWork
tugrazLancaster University
Causality AnalysisFirst Stage: Bayesian Longitudinal Analysis
Correlation Structure: Latent abilities on the same subject
will be more correlated than among different subjects.
1 Autoregressive AR(1)
Constant variance across time.
Correlation exponential decrease as the lag between
times increases.
Σθ = σ2
1 ρ ρ2
ρ 1 ρρ2 ρ 1
2 Unstructured Covariance: Not specific pattern
3 Random Effects: θift = γ(0)if + γ
(1)if t
3 chains of 10000 iterations with a burn-in phase of 5000 and
final results pooled in a single chain.
Employment of a BUGS (Bayesian inference Using Gibbs
Sampling) code called from the free software R.
Vilma Romero VI COBAL8 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
MIRT
Causality Analysis
Conclusionsand FutureWork
tugrazLancaster University
Prior Distributions
Choice of prior distributions for each f latent dimension
Summary of DIC criterion
Model DIC Q0.025 Q0.975
AR(1) Covariance Structure 2312.46 2205.88 2418.96Unstructured Covariance 2242.62 2124.69 2359.80Random Effects 2337.56 2258.15 2415.93
Vilma Romero VI COBAL9 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
MIRT
Causality Analysis
Conclusionsand FutureWork
tugrazLancaster University
Estimation Results - AR(1)Summary of ρ estimate
−2
−1
0
1
2
0
1
2
3
Difficulty
Discrim
ination
NARTESQPRETENSE
DESIRENVFBTR1
DBSELFQNVFBTR3
NVFBTR2NVFBTR4
THINKDBOTHEQ
VFBQ1T1VFBQ1T2
STANTES
Credibility Interval of Item parameters
Vilma Romero VI COBAL10 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
MIRT
Causality Analysis
Conclusionsand FutureWork
tugrazLancaster University
Estimation Results - AR(1)
●
●●
●●
●● ●
●
●● ●
● ● ●
●
●●
●
●●
●
●
●
● ●
●●
● ●●
● ●●
●●
●
●
●
● ●
●
● ●
●
●● ●
●
● ●
●
●●
●
●
●
●
●
●
●
●●
●
●
●●
● ●●
● ●
●
●
●
●
●
●
●●
●
● ●
●● ●●
●
●
●
● ●
●
●
●●
●
●●
●
●●
●
● ●
●
●● ●
●
●
●
● ●
●●
●
●
● ●
●●
●●
●●
●
●
●
●●
●
●
●
●●●
●●
● ●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●● ●●
●●
●●
●●
● ●● ●
●
●
●●
●● ●
●
●●
●
●●
● ●
●
●
●
●
●
●
●
●● ●
●
●●
●●
●
●
●
●
● ●
●●
●
●
● ●●
●
●●
●
●
●
●●
●
●
●
●
● ●
●● ● ●
●
●●
●
●
●
●●
●
●●
●
●
●
●
●●
●●
● ●
●●
●
●●
●
● ●
●
● ●●
●●
●●
● ●
●
●
●
● ●
●●
●
●
●● ●
●● ●
●
●●
●●
●
●
●●
●●
●
●
●
●● ● ●
●
●●
●●
●●
●●
●●
●●●
●●● ●
●
●●
●
● ●●
● ●●
●●
●
●
●
● ● ●
●
●●
●
●
●
●
●●● ●
●
●
●
●
● ●●
●
●
●
● ●
●●
●
●
●
●
●
●● ●
●● ●
●
●●
●
●
●
● ●
●
●●
●
● ●
●● ● ●●
●●
●
●
●
●●
●
●
●
●
● ●
●●
●●
●
●
●● ●
●
●●
●
●●
●
●
●
●●
● ●●
●●
●●
●
●●
●● ●
●
●
●●
●● ●
●
●●
●
●
●
●
●
●
●
●
●
●
● ●
● ●●●
●●
● ● ●●
●
●
● ●
●
●● ●
● ● ●●
● ●
●
●
●
● ●
●
● ●
●
● ● ●●
● ●●
●
●
● ●
●● ●
●
●
●●
●● ●
●● ●
●
●●
●●
●
●
●●
● ●
●●
●●
●● ●
● ●●
●
●
●
●
●
●
● ●
●
●● ●
●●
●
●
●
●● ●
●
● ●
●
● ●
●
●● ●
●●
●
●
●●
●
●●
● ●
●●
●
●●
●●
●● ●
●
●●
●
●
●
●
●
●
●
●
●●
●●
●● ●
●
●
●
● ● ●
●
●
●
●
●
●
● ●
●●
● ●
●
●●
●●
●●●
●
● ●
●
● ●
●
●● ●
●
●●
●●
●
●
●
●
●●
●
●● ●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●● ●
●
●● ●
●
●●
●●
●●
●
●
●
●
●
●● ●
●●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●● ●
●
●●
● ●●
●
●
●
● ●
●
●●
●●
● ●
●
●●
●
●
●
●
●
●● ●
●●
●●
●● ●
● ●●
● ●●
●
●●
●
●
●
●● ●
●
● ●●
●●
●
●
●
●●
●
● ●
●
●● ●
●● ●
●
●●
● ●
●
●
●●
●●
●
●
●●
●● ●
●
●●
●
●
●
●
●
●
● ●
●●
●●
●● ●●
●●
●
●
●
●
●
●
●●
●
●●
●
●
● ●●
● ●
●
●
●
●●
●
●
●
●
●
●
●
●● ●
●
●
●
●
●
●
●●
●
●●
●
●
●●
●
● ●
●
●●
●
●
●
●●
●
● ●●
● ● ●●
● ●
●●
●
● ●
●
●●
●
●
●
●
●● ●
●
● ●
●
●● ●
●●
● ●
●●
●●
●● ●
●
● ●
●
●●
●●
●
● ●
●
●
●●
●
●
●
●● ●● ●
●
●●
●●●
●●
●
●
●● ●
● ● ●●
● ●
●
●
●
●●
●●
●
●
●● ●
●
● ●●
● ●
●
●
●
●
●
●
●●
●
●● ●
●● ●
●
● ●
●
●
●
●
●
●
●●
●
●
●●
● ●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
● ●●
●
●●
●
●
●
●
●
●
● ●
●
●● ●
●
● ●
●
●●
● ●
●
●●
●
● ●●
●
●
●● ● ●
●
●●
●●
●
● ●
●
●
●
●
●● ●
●
● ●
●
●
●
● ● ●
●
●●
●
●
●
●● ●
●●
●
●
●● ●
●
●
●●
●
●
●
●
●● ●
● ●●
●
●●
●
●
●
●● ●
●
●
●
●● ●
●● ●
●
●●
●
●●
●●
●
●
●●
●
●
●
●● ●
●●
●
●
●●
●
●
●
● ●
●
●● ●
●●
●
●
●●
1 2 3 4 5 6 7 8 9 10 11
12 13 14 15 16 17 18 19 20 21 22
23 24 25 26 27 28 29 30 31 32 33
34 35 36 37 38 39 40 41 42 43 44
45 46 47 48 49 50 51 52 53 54 55
56 57 58 59 60 61 62 63 64 65 66
−1
0
1
2
−1
0
1
2
−1
0
1
2
−1
0
1
2
−1
0
1
2
−1
0
1
2
1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
Time
Late
nt A
bilit
y
● ● ● ● ● ●Non Verbal FB Pretense, Desire, Think Verbal FB Deceptive Box Narrative Location Change
Estimated Latent Ability by subject
Vilma Romero VI COBAL11 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
MIRT
Causality Analysis
Conclusionsand FutureWork
tugrazLancaster University
Second Stage: Ability Regression
Regression of the latent ability factors of t = 2, 3 against the latent
ability of the previous instant of times.
●●
●●
●●
●
●
●
●●●
●
●●
●●
●
●●Location Change
Narrative
Deceptive Box
Verbal FB
Pretense, Desire, Think
Non Verbal FB
1 2 3Time
0.01
0.02
0.03
0.04
p−value
R2
●
●
●
0.4
0.6
0.8
Path Diagram of Causality - Model AR(1).The p-values have not been adjusted for multiple comparison.
Vilma Romero VI COBAL12 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
Conclusionsand FutureWork
tugrazLancaster University
Conclusions
1 Children before 4 years old successfully passed Pretense,Desire and NVFB tasks.
2 ToM reduced to 6 latent abilities through the Bifactor Model.
3 Easy items: Pretense and Desire.
Most difficult item: Standard Location Change.
4 Significant improvement across time: NVFB ability.
5 Causal analysis: Pretense, Desire and Think affects the
development of most of the others abilities.
Future Work
Consider the correlation between latent abilities in the model.
Include a guessing parameter for each item.
Covariates (age, sex and institution) could be included.
Multilevel Modelling or Dynamic Latent Trait Models.
Vilma Romero VI COBAL13 / 14
Introduction
DataDescription
ExploratoryAnalysis
Results
Conclusionsand FutureWork
tugrazLancaster University
Thank you!!!
Vilma Romero VI COBAL14 / 14