Please enable JavaScript.

Coggle requires JavaScript to display documents.

ITEM ANALYSIS, A positive item-total correlation - Coggle Diagram

- - - - method of quantifying these differences
        
        When nobody choose the correct answer (p value = 0.0)
        
        When everyone choose the correct answer (p value = 1.0)
        
        There are no individual differences
- - - - If a test is given to a large group of people - comparing
        
        High test scores (e.g. those in the top 25% of the class) who answered that item correctly.
        
        Low scores (e.g. bottom 25%) who answered the same item correctly.
      - The steps in computing item discrimination index (D)
        
        Select an upper group and a lower group
        
        Calculate the percentage
    - - a statistic that provides
        information
        
        on the degree to which a particular item
        
        total test measure the same thing
      - This statistic represents the correlation between
        
        the score on
        an item
        
        the total test score
      - An item-total correlation near zero
        
        The item does not discriminate between high and low scores
      - A negative item-total correlation
        
        The scores on the item and scores on the total test disagree
      - advantage of the item-total correlation
        
        It is the simple correlation between
        
        item scores
        
        test score
        
        easy to make judgment about practical significance
        
        If the item-total r = .40
        
        the item can account for 16% of the
        variability in test scores.
        
        An item discrimination index of .40
        
        cannot be re-expressed in
        such concrete terms.
        
        Item-total correlations are directly related to the reliability
    - - If the item-total correlation is low
        
        The item in question is not correlated with any of the
        other items on the test.
        
        The item should be rewritten or discarded altogether
        
        The item shows positive correlations with some test but shows correlation near zero, or negative correlations
        with other items on the test.
        
        Occur if the test measured two distinct attributes
      - The result of inter item correlation matrix is packed with information
        
        To compute the reliability
        
        the average inter item correlation and the number of items on the test
        
        To understand the measures of item discrimination