2016-07-07-Teleconference

[andkov]

• analytic starters (seeds, seedlings) - meta, and individual study
• clarify (specifically) the intended use
• What words to add in the seed?

  - how to interpret this table?

• drivers need THE model-level info

  • variances of the intercept, slope,
  • the size of the variance of the slope will affect the significance of the correlation. Needs to be addressed before we can safely interpret the pattern
  • value
  • significance
  • meaning/interpretation

    series of tests to rule out concerns:

  We don't want to report the results of the item level (eg. logic_im, bstory_im), but on on the DOMAIN level.

  • process and domain ARE NOT THE SAME!
• how is BIVARIATE growth reported? see examples in the literature. The growth pro

Procedure of drawing inferences:

1. Is the correlation b/w slopes significance
   • build the branching logic for moving through evidence for the existence of the trend
   • should DRIVERS
   • dynamic - to explore
   • static - to focus attention

     Cosmetics

   • order the female

     Variance of slope

   • variances of slopes: what to do about them?
   • deal with dynamic tables or rearrange (need to make more convenient)
   • add variance components to he growth curve table

     Analytic sequence

   • 1. correlation b/w slopes
   • 2. variance of slopes
   • 3. ...

        Keep domains in mind

        - item is need for consistency

[andkov]

[andkov]

[andkov]

Focusing on a physical process

a similar table will be produced for each physical process in the given study

[andkov]

Focusing on a cognitive process

a similar table will be produced for each cognitive process in the given study

[andkov]

Model look up

library("magrittr")
results <- read.csv("https://raw.githubusercontent.com/IALSA/IALSA-2015-Portland/master/data/shared/derived/temp.csv", stringsAsFactors = FALSE)
source("https://raw.githubusercontent.com/IALSA/IALSA-2015-Portland/574bc57a0375d41bf267e9808e5d77f5473832b8/scripts/model-lookup-function.R")

apply the function

single_model <- pull_one_model(d = results,
                                      study_name_ = "map",
                                      subgroup_   = "female",
                                      process_a_  = "grip",
                                      process_b_  = "bnt",
                                      model_type_ = "aehplus"
)
print(single_model)