ECED 3209 Lecture 5: Week 5 Assessment
Document Summary
Split test in half and administer halves separately. Look at correlation between between two halves. If all students show a high correlation, your test is likely reliable. Same test to same person on two different occasions. Shorter time gap= more likely scores will match. Reliable if a longer time gap and scores match. True score: the score a person would receive if the construct could be perfectly measured. Observed scores: the score a person actually receives on a measure. Measurement error: other characteristics, besides the true score, that influence the observed score. Large set of questions that address the same material and then divide the questions into two sets. Correlation between scores on each parallel assessment. You want it to be no matter which test form they take they get the same test score. Used to assess the consistency of results across items within a test- do each of the questions on a test correlate with one another.