We will perform four separate analyses: (1) for experiments with visual noise in younger and older adults; (2) for experiments with auditory noise in younger and older adults; (3) for experiments with visual noise in young adults and adolescents; (4) and for experiments with auditory noise in young adults and adolescents. For each of the four analyses, we will build two linear regression models: one with question response accuracy as a dependent variable and the other one with mean per word reading times of the stimuli sentences as a dependent variable. Analyses (1) and (2) will differ in the type of noise we will include in the models, but the structure of the models will be the same. The first logistic regression model will include accuracy as a dependent variable and seven independent variables (plausibility, closure, noise, plausibility * noise, plausibility * age, age * noise, plausibility * age * noise), as well as by-subject and by-item random intercepts. We will also include by-subject random slopes for plausibility, closure, noise, and plausibility * noise as well as by-item random slopes for plausibility, closure, noise, plausibility * noise, plausibility * age, age * noise, plausibility * age * noise if the model with this random effect structure converges. Independent variables will be coded with sum contrasts (contr.sum) with two levels each (visual or auditory noise / no-noise; younger age / older age; plausible / implausible; high attachment / low attachment). After that we will create a model matrix (model.matrix(~ Plausibility * Age * Noise + Closure, data=data)) and extract the particular contrasts as independent variables. glmer(Accuracy ~ 1 + Plausibility + Closure + Noise + Plausibility_Noise + Plausibility_Age + Age_Noise + Plausibility_Age_Noise + (1 + Plausibility + Closure + Noise + Plausibility_Noise | Subject) + (1 + Plausibility + Closure + Noise + Plausibility_Noise + Plausibility_Age + Age_Noise + Plausibility_Age_Noise | Item), data = data, family = binomial) The second linear regression model will include log-transformed reading times as a dependent variable and four independent variables (noise, age, plausibility * noise, age * noise), as well as by-subject and by-item random intercepts. We will also include by-subject random slopes for noise and plausibility * noise as well as by-item random slopes for noise, age, plausibility * noise, and age * noise if the model with this random effect structure converges. Independent variables will be coded with sum contrasts (contr.sum) with two levels each (visual or auditory noise / no-noise; younger age / older age; plausible / implausible). After that we will create a model matrix (model.matrix(~ Plausibility * Age * Noise, data=data)) and extract the particular contrasts as independent variables. lmer(RT ~ 1 + Noise + Age + Plausibility_Noise + Age_Noise + (1 + Noise + Plausibility_Noise | Subject) + (1 + Noise + Age + Plausibility_Noise + Age_Noise | Item), data = data, REML=FALSE) ___________ Analyses (3) and (4) will be similar to the analyses we described above, with three changes in the fixed effects structure and two changes in the random effects structure in the regression models. The first model will include age as an independent variable and will not include the interaction of plausibility * age * noise. Overall, this model will comprise seven independent variables: plausibility, closure, noise, age, plausibility * noise, plausibility * age, age * noise. It will also have by-item random slopes for age. glmer(Accuracy ~ 1 + Plausibility + Closure + Noise + Age + Plausibility_Noise + Plausibility_Age + Age_Noise + (1 + Plausibility + Closure + Noise + Plausibility_Noise | Subject) + (1 + Plausibility + Closure + Noise + Age + Plausibility_Noise + Plausibility_Age + Age_Noise | Item), data = data, family = binomial) The second model will include the interaction of plausibility * age as an independent variable and overall will have five independent variables: noise, age, plausibility * noise, plausibility * age, age * noise. It will also have by-item random slopes for plausibility * age. lmer(RT ~ 1 + Noise + Age + Plausibility_Noise + Plausibility_Age + Age_Noise + (1 + Noise + Plausibility_Noise | Subject) + (1 + Noise + Age + Plausibility_Noise + Plausibility_Age + Age_Noise | Item), data = data, REML=FALSE)