Session Ready
Exercise

Internal reliability

You know how to examine how individual items perform in your measure, but what about how well those items relate to each other - the overall internal reliability of a measure? Coefficient alpha (also called Cronbach's alpha) and split-half reliability are two common ways of assessing reliability. These statistics are a function of the measure length and items' interrelatedness, which you just investigated by looking at the correlation matrix.

In reliability values greater than 0.8 are desired, though some fields of study have higher or lower guidelines.

Do the results of alpha() and splitHalf() indicate the gcbs dataset has acceptable reliability?

Instructions 1/3
undefined XP
  • 1
  • 2
  • 3
  • Use alpha() to get estimates of coefficient alpha and inspect the results. Be sure to check out the information about how reliability would change if items were dropped.