Lec. 5 Correlational Research

Question 1

Correlation

Answer 1

CORELLATIONAL RESEARCH – searches for a connection between scores, whether two variables are related to one another.

• Ex: The teacher is interested in whether there is some connection between the amount of time students spend in preparation for a test and the test scores those students get.
• RESEARCH HYPOTHESIS: “As preparation time increases, test scores tend to increase.”

CORRELATIONS:

• Can either be positive or negative.
  
  • Positive: As scores in one variable increase, scores in the other variable tend to INCREASE.
  • Negative: As scores in one variable increase, scores in the other variable tend to DECREASE.
    
    • more cigarettes a person smokes, the shorter his/her life span is likely to be

CORRELATION COEFFICIENT (r) – Correlation is measured with correlation coefficient r – reveals how closely two things vary together and how well one predicts the other.

• Knowing how much SAT scores correlate with college success tells us how well the scores predict college success
• Values of r fall within the range from –1.00 to +1.00
  
  • HIGH +/- .7 - .9
  • MODERATE +/- .5 - .6
  • LOW +/- .4 - .3
  • NONE +/- .1 - .2
• Absolute value is close to +1.00 is called a high correlation.
• Signed value is close to +1.00 would be called a high POSITIVE correlation.
• Signed value is close to –1.00 would be called a high NEGATIVE correlation.
• ****************Correlation does not mean causation!

Question 2

3rd Variable Problem

Answer 2

THIRD VARIABLE PROBLEM (CONFOUNDING VARIABLE) – Correlation simply indicates a relationship – although correlation tells us nothing about cause and effect, it reveals the actual extent to which two things relate. The actual CAUSE of the relationship could be one of three things:

• A causes B
• B causes A
• Some THIRD VARIABLE (C) causes both A and B
  
  • Ex: Scientists have linked television-watching with childhood obesity.
    
    • One explanation is that TV watching results in less exercise and more snacking (A causes B).
    • Another explanation is that obesity may lead children to prefer more sedentary activities, such as TV viewing (B causes A).
    • Or, perhaps, some third factor may explain the relationship. For example, parents having little formal education may not emphasize good nutrition or good use of leisure time (C causes both A and B).
  • Ex: Children with high self-esteem also tend to have high academic achievement.
    
    • Some might argue that a healthy self-concept boosts school achievement.
    • Others are convinced that high achievement produces a favorable self-image.
    • Reports of a recent nationwide sample indicate neither is true. In other words, self-esteem and achievement are not causally connected. Rather, their correlation is due to their both being linked to intelligence and family social status.
      
      • Remove the effect of these two variables, and the correlation between self-esteem and achievement evaporates.
  • Ex: A Dartmouth study found that overweight young women (age 23) earned 6.4% less than their non-overweight peers. Additionally, the study found that young men’s earnings rose 2% for each 4-inch increase in height.
    
    • Is this study correlational or experimental? Correlational.
    • If this study is correlational, how are the variables are related? Is the correlation positive or negative? What is the possible third variable? The weight study is a negative correlation (as weight went up, earnings went down). The height study is a positive correlation (as height increased, earnings increased).
  • Ex: An Australian study reported that MSG does not cause people to be sick, as previously reported. The researcher told subjects that he was studying ingredients in a new soft drink and fed them either MSG or a placebo in the drink. The same number and type of symptoms were reported in both the MSG and the placebo groups.
    
    • Is this study correlational or experimental? Experimental.
    • If this study is experimental, how is variable A is causing variable B? MSG or placebo is the independent variable. Symptoms are the dependent variable. The manipulation of the MSG causes the outcome, which shows no difference in symptoms between MSG and the placebo.