0
$\begingroup$

I conducted a log linear analysis on a dataset where i have 3 categorical columns to compare so i can see if there is an effect of on others or a combination of two ....

the description is here :

data info

Then i get this result :

enter image description here

But I am still not sure if my Assumptions are violated, and if is the case is there an alternative test that i can use.

My main objective is to define how quest_Type and/or Lenght_Str effect on isCorrect.

3: parameter estimation

The test i used :

enter image description here

And here are the variables i used :

enter image description here

Improve this question
$\endgroup$
8
  • 7
    $\begingroup$ You have 1,000,000 missing cases? and 8000 "out of range"? And 47 df with only ~300 cases? And lots of 0 chisquare? Why? $\endgroup$
    – Peter Flom
    Commented Dec 7 at 0:35
  • $\begingroup$ @PeterFlom maybe the number of iteration (99) ? i edited the image result i'm using now 10 iteration . $\endgroup$
    – MexcelsiorB
    Commented Dec 7 at 11:39
  • 1
    $\begingroup$ If you have 1,000,000 missing cases, and 8,000 "out of range", and only 304 valid cases, then the number of iterations is not your first problem. Please describe your data. $\endgroup$
    – Stephan Kolassa
    Commented Dec 7 at 19:17
  • $\begingroup$ @StephanKolassa here is my data : docs.google.com/spreadsheets/d/… , i just remove the modality "baseline" from question_type column. $\endgroup$
    – MexcelsiorB
    Commented Dec 7 at 19:34
  • $\begingroup$ @StephanKolassa my main goal is to decide if there is an effect of question_type and/or sting_length columns on isAnswercorrect column. The three columns are categorical variables. question_type = 4 modalities , sting_length= 4 modalities , isAnswercorrect = 2 modalities. $\endgroup$
    – MexcelsiorB
    Commented Dec 7 at 19:36

0

Reset to default