RUMM2030
For analysing assessment and attitude questionaire data


(click on images to enlarge)

 

 

 

 

Polytomous responses
  • consistent estimates of item parameters;

  • rating scales.

  • "partial credit";

  • person and item fit;

  • threshold analysis (illustrated);

  • quantifies effect of local dependence.

Rescore

  • change scoring function for items presenting non-disctiminating categories;

  • form "partial credit" from multiple choice items using distractor analysis.

Differential Item Function [DIF]

  • specify person factors, eg. gender, country;

  • specify levels, eg. male, female;
  • each level plotted against the ICC (illustrated);
  • does ANOVA of person factor against class interval.

Measure

  • estimate person location from calibrated test items;

  • item estimates imported [templates] or selected from project;
  • use all items or select subset only;
  • large record numbers possible.

Equating Test Score

  • select subset of items from item pool;

  • plot ICC for each subset;
  • select equating criteria;
  • read equivalent values from plots (illustrated);
  • display t-test analysis for subset-pairs.

Other Features

  • combine multiple choice and polytomous items in the one analysis;

  • simple visual controls for selecting data file and response format;
  • amend illegal reponses, eg. typing errors in response data file;
  • allocate illegal responses as missing data;
  • distractor analysis for MC items to checking for cheating and strength/effectiveness of response distractors across an item;
  • create item subsets for assessing level of dimensionality, local dependence;
  • create complete data sets by deleting persons with missing data.