On my latest physician survey report I noticed
that administrator skill was at the top of the list of questions
that are correlated with overall satisfaction.  The report said that
the correlation between administrator skill and the overall satisfaction of physicians was .80.  Does that mean that high administrator skill causes physicians to be satisfied?

Response:

That is a great question!  It is a common belief that if two things are correlated, one causes the other.  This is actually far from the truth and is a basic fallacy in human critical thinking.  If two things occur at the same time, we tend to think that one causes the other. 

For example, a hospital employee notices a beautiful, full moon on the way to work the night shift.  Once at the hospital, the employee perceives that patient behavior is unusually bizarre.  The basic human tendency is to connect the two events and conclude that a full moon causes disturbed behavior in the hospital.  In actuality, it is impossible to conclude that the full moon is causing the behavioral disturbance.  In reality, other factors could explain the behavior in the hospital. 

A similar situation occurs with your physician satisfaction results.  Your physician report identifies the correlation or “relationship” between your physicians’ overall satisfaction with the hospital and other survey questions.  These relationships are summarized by what is called a “correlation coefficient” that ranges from -1 to + 1.  The closer a correlation coefficient is to +1, the more the two items are positively correlated or related.  In other words, as scores on one item increase so do scores on the other item.  Oppositely, the closer the correlation coefficient is to -1, the more the variables are related but in opposite directions.  As scores on one variable increase, scores on the other variable decrease. 

As you interpret the correlation of +.80 in the example above, it is important to remember that administrator skill was highly related to physician overall satisfaction with the hospital—but we cannot conclude that the skill level of administrators caused the physicians to be satisfied.  The scores on these two items were collected at the same time by the same physicians but we must be careful not to draw an erroneous conclusion.  It may be that other variables are influencing the satisfaction of physicians.  For example, administrator communication with physicians and administrator responsiveness also tend to be highly correlated with the overall satisfaction of physicians.  Thus, all three of these items may be influencing the satisfaction of physicians.  We simply cannot assume that any of them “causes” physician satisfaction to go up or down. 

Items on satisfaction surveys are often “inter-correlated.”  Therefore, improvement in one area could influence scores in other areas.  As you think about improving the level of satisfaction of physicians in the hospital, you might want to work on several of the most highly correlated items with overall satisfaction.  Intervening in areas that correlate highly with physician satisfaction on which you score low can also be a valuable strategy.   

Tips to remember about correlation coefficients:

• The range of the correlation coefficient is -1 to +1
• The closer the coefficient is to either -1 or +1, the stronger the relationship
• The closer the coefficient is to zero, the weaker the relationship
• A positive correlation means that as one variable increases, so does the other
• A negative correlation means that as one variable increases, the other decreases
• Correlations between .2 and .3 (either positive or negative) are weak in strength
• Correlations between .4 and .6 (either positive or negative) are moderate in strength
• Correlations from .7 and up (either positive or negative) are strong
• Correlation does not imply causality, but suggests relationship