Research Methods 1. 2. 3. Case studies Correlational research Experimental research **will need to know more detail than is in your book about these topics Case Study Murray (1938) personology = need to understand the whole person as a coherent entity (vs. just parts of people) in-depth study of one person Case Study Useful in several instances: 1. Rare or unusual situations 2. To demonstrate possibilities 3. Hypnosis To demonstrate a treatment 4. Piano virtuoso Mass murderers PTSD treatment in child As disconfirming evidence Shatter assumptions about abilities Case Study Strengths Depth and detail Capture complexity Weaknesses Problem of generalization: idiosyncratic subject Problem of generalization: experimenter biases and subjective impressions Entangled relationships among variables Correlational Design = a study that assesses the extent to which two variables are related Defines the relationship in quantitative terms Correlational (“co-related”) When one variable changes in value, what happens to the other variable? Correlation Example Is there a relationship between self-esteem and GPA? 1. Need to have different levels of my first variable: self-esteem Very high self-esteem -------- ? Moderately high self-esteem--? Average self-esteem -----------? Moderately low self-esteem --? Very low self-esteem ----------? Correlation Example Raw Data: Self-esteem score Tim 42 Bart 10 Kelsey 15 Kim 22 Etc. GPA 3.8 1.4 2.5 3.1 Correlation Example See scatterplot of data Self-esteem and GPA data 4 3.5 3 GPA 2.5 2 Series1 1.5 1 0.5 0 0 5 10 15 20 25 self-esteem 30 35 40 45 Direction of Correlation Scatterplot showed a positive correlation As one variable increased, the second variable also increased As self-esteem goes up, academic achievement also goes up Think of some examples of positively correlated variables Negative (inverse) correlation As on variable increases, the second variable decreases (i.e. one gets bigger, the other gets smaller) As amount of alcohol intake increases, motor control decreases Think of examples of negatively correlated variables = direction of the correlation Strength of Correlation How strongly related are the two variables of interest? the “sloppiness” of association Degree of accuracy with which you can make a prediction about 2nd variable given value of the first variable Ranges from -1 to 1 -1 and 1 are very strong (perfect) correlations 0 is no correlation; no relationship Correlation – strength and direction Correlation Example High Self-esteem and GPA Does (A) lead to (B)? Or is the other way around? Or, are there other factors that lead to both (A) and (B)? Two independent carefully conducted studies found that there is no causal relationship between these two factors. They are correlated because both of them are correlated to some other factors: intelligence and family social status. **Correlations do NOT tell us that one variable CAUSES the other variable. A recent Morgan-Gallup Poll in the US of 1009 people asked: “Does correlation imply causation?” 64% YES 38% NO 8% undecided Correlational research Strengths Can study a broad range of variables Can look at multiple variables at one time Large samples are easily obtained Weaknesses Relationships established are associational, not causal Individuals not studies in-depth Potential problems with reliability and validity of selfreport measures Experimental Design Allows us to determine cause and effect Defining characteristics: 1. Manipulation of variables 2. 3. Independent variable Dependent variable Experimental control of other variables Random assignment to groups Example Learned helplessness All subjects first hear a very loud noise 3 groups: Can end the noise by pushing a button Cannot stop the noise Control group – doesn’t hear noise Put in 2nd situation where they could end a loud noise by moving their hand. D.V. was response latency (how soon did they move hand?) 1st and 3rd groups learned quickly to move hand; 2nd group sat passively and did nothing Experimental Design Strengths Can tease out cause and effect Allows for strict control of variables Weaknesses Many questions may not be able to be answered using this method – i.v. cannot be varied (e.g. sex, age, birth order, effects of child abuse) May be artificial and limited Causal effects may not hold when the complexity of actual human behavior is considered Involves brief exposures and may miss important processes that occur over time Example from the news/class activity EXAMPLE 1: "MARRIAGE SLOWS CANCER DEATHS" Evidence that married people have a better chance of surviving cancer than do singles means that the unmarried might be good targets for cancer-prevention programs. Married people with cancer had a 23% higher overall survival rate than the unmarried. Another example Example 2: Children who are aggressive tend to watch a higher proportion of violent television than children who are not highly violent or aggressive What type of design? What can we conclude? Categorize I give students a questionnaire that measures how much they like sensation seeking activities. I then ask them about current drug use. I find that students who are high in sensation seeking engage in more drug use than students low in sensation seeking. What type of study? What type of relationship? Categorize I randomly assign children with behavior problems to two types of play groups: one group is structured and organized and the other group has no rules. I measure the number of aggressive behaviors in each group. What type of research is this? What are the independent and dependent variables? Does the type of research change if I look at children with behavior problems and compare them to children without behavior problems? Categorize Over the course of several years, I interview three adolescents who live in poverty. I am interested in the impact that poverty has on their lives. What type of study?