search

GaboSJW / online_presentation

0 stars 0 forks source link

Presentation Comments #1

Open knf43 opened 3 years ago

knf43 commented 3 years ago

Hi Jiawei, I have a few questions about your presentation:

  1. Were the assumptions for the statistical tests conducted mentioned in the paper? For example, was the data normally distributed (an assumption of t-tests)?
  2. For the GLMM analysis, could you clarify what you mean by "the purpose as to test if the fixation time can be fit as a function of participant group, syllable structure, stress, and WM?" Based on the research questions, it seems like the authors were interested in testing the effect of syllable structure, stress, and working memory (the independent variables) on fixation time (the dependent variable) across groups (SS, LA, and LB). Is this right? Also, on a related note, was it mentioned why the authors ran generalized linear fixed-effects models (GLMM)? In other words, why did the authors choose this type of analysis based on the data?
  3. On slide 10, you mentioned that the authors mentioned a marginally significant result (i.e., p = .03) for the lower beginner group and paroxytone targets. However, when I look at the table on slide 7, it looks like there is another marginally significant result (i.e., p = .02) for the advanced learners and oxytone targets. Was this mentioned in the paper? If not, why do you think one marginally significant result was noted and not the other? Thank you! :-)
GaboSJW commented 3 years ago

Hi Jiawei, I have a few questions about your presentation:

  1. Were the assumptions for the statistical tests conducted mentioned in the paper? For example, was the data normally distributed (an assumption of t-tests)?
  2. For the GLMM analysis, could you clarify what you mean by "the purpose as to test if the fixation time can be fit as a function of participant group, syllable structure, stress, and WM?" Based on the research questions, it seems like the authors were interested in testing the effect of syllable structure, stress, and working memory (the independent variables) on fixation time (the dependent variable) across groups (SS, LA, and LB). Is this right? Also, on a related note, was it mentioned why the authors ran generalized linear fixed-effects models (GLMM)? In other words, why did the authors choose this type of analysis based on the data?
  3. On slide 10, you mentioned that the authors mentioned a marginally significant result (i.e., p = .03) for the lower beginner group and paroxytone targets. However, when I look at the table on slide 7, it looks like there is another marginally significant result (i.e., p = .02) for the advanced learners and oxytone targets. Was this mentioned in the paper? If not, why do you think one marginally significant result was noted and not the other? Thank you! :-)

Hi, thank you for your questions.

As far as I can tell, there was no explicitly mentioned if the data were normally distributed or not. Your interpretation is 100% correct. Sorry if my explanation wasn't as clear. In the paper, the authors explained the reason why GLMM was used as it was best fit when including the random effects. In statistics, a generalized linear mixed model (GLMM) is an extension to the generalized linear model (GLM) in which the linear predictor contains random effects in addition to the usual fixed effects. It was not mentioned in the paper. The marginally significant result on slide 10 was mentioned because of its inconsistency with the result of the other groups(normally paroxytone < oxytone, but in LB group was paroxytone > oxytone). In other words, it was mentioned as a surprising result but not because of the level of significance. On slide 7, the resulting pattern was consistent across all groups, plus other insignificant results are reported as >0.008, which could be smaller than 0.02, so the 0.02 might not be marginally significant in this analysis.

knf43 commented 3 years ago

Hi Jiawei, Got it, thank you so much for clarifying!