Assessment, Statistics, and Research

  In research, alpha is  the probability of rejecting a true null hypothesis.  In testing, alpha also refers to a measure of internal consistency—see Cronbach’s coefficient alpha. Alpha waves are brain waves that can be measured on an electroencephalogram (EEG). Alpha waves are associated with daydreaming, meditating, and mindfulness. Applied Statistics Concepts for Counselors on    AMAZON  or   GOOGLE Please check out my website     www.suttong.com    and see my books on    AMAZON        or  GOOGLE STORE Also, consider connecting with me on    FACEBOOK     Geoff W. Sutton         TWITTER    @Geoff.W.Sutton     You can read many published articles at no charge:   Academia    Geoff W Sutton      ResearchGate    Geoffrey W Sutton   Please check out my website     www.suttong.com    and see my books on    AMAZON        or  GOOGLE STORE Also, consider connecting with me on    FACEBOOK     Geoff W. Sutton         TWITTER    @Geoff.W.Sutton     You can read many published artic

  Coefficient Alpha (also called "alpha") is a statistical value indicating the degree of internal consistency of items in a multiple-item scale like survey items or Likert-type scales. Internal consistency is one measure of reliability for scores from scales, measures, and survey items. The alpha statistic was developed by Lee Cronbach in 1951 thus it is also called Cronbach's alpha . In research reports, you may just see the Greek lower case letter alpha,  α. The procedure to calculate alpha can be found in SPSS under Analyze > Scale > Reliabilty. For research purposes, scales with alpha levels equal to or above alpha = .70 are acceptable. The best scales have values of alpha = .9 or higher. The alpha method works best to evaluate unidimensional measures. If there are two or more dimensions in a set of items, the alpha value will be lower so, when alpha values are low, consider which item or items do not support the primary dimension. Cite this Post Sutton, G.W.

Why aren't tests reliable? The reason tests are not reliable is that reliability is a property of the interpretation of scores not the tests themselves.   This isn't a matter of semantics. Think about it this way. Give all the students in one school an achievement test. The test items don't change so they appear stable, consistent, and reliable. However, when publishers report reliability values, they calculate the reliability statistics based on scores. Scores vary from one administration to another. If you ever took a test twice and got a different score, you know what I mean. Individuals change from day to day. And we change from year to year. Also, even a representative sample of students for a nation can be different each year. Every time we calculate a reliability statistic, the statistic is slightly different. Reliability values vary with the sample. Reliability values vary with the method of calculation. Reliability values also vary with the method us