Presentation on theme: "The Psychology of the Person Chapter 2 Research Naomi Wagner, Ph.D Lecture Outlines Based on Burger, 8 th edition."— Presentation transcript:
The Psychology of the Person Chapter 2 Research Naomi Wagner, Ph.D Lecture Outlines Based on Burger, 8 th edition
The Hypothesis-Testing Approach Research begins with a theory, which is a general statement about the possible relationship between variable (explaining) Theories differ in the range of events or behaviors they attempt to explain A good theory is parsimonious- uses as few variables as possible, and useful- lends itself to research.
Theories and Hypotheses A hypothesis is a specific prediction that is logically derived from the theory, and lends itself to be tested in an experiment A hypothesis is a specific prediction that is logically derived from the theory, and lends itself to be tested in an experiment The theory itself is never tested directly and cannot be proven or unproven It can be supported by research A good research progresses fro theory, to hypothesis (prediction) to experiment. The experimental variables: The Independent variable- supposed to have an impact on the dependent variable
The Experiment A procedure that enables us to identify cause-and effect relationships We manipulate the Independent Variable (IV) to assess its influence on the Dependent Variable (DV) To Manipulate the IV means to create different levels of it, and assign people to be exposed to the levels For example: If we want to assess the impact of sleep deprivation on concentration, we will create different levels of sleep- to see whether those who sleep more are able to concentrate better than those who sleep less ( note issue of random assignment)
An Example Theory: a connection between sleep and the ability to focus in the classroom Hypothesis: Those who sleep more are better able to focus The Independent variable: The amount of sleep The dependent variable: the ability to focus
The Experimental Variables We manipulate the IV by forming different levels of it, for example, we assign participants into sleep groups We manipulate the IV by forming different levels of it, for example, we assign participants into sleep groups 8 hours sleep 5 hours sleep 3 hours sleep 0 hours sleep We want to see whether those who sleep more were better able to focus (work on a reading task) than those who sle
How can we eliminate the impact of other factors on the ability to focus ? There might be other factors beside sleep that may affect the ability to focus- for example, someone is sick, does not speak the language, has just now got bad news Random assignment of participants into the various experimental groups (e.g. sleeping 8 hours, 5 hours, 3 hours etc) eliminates any pre-existing differences among the participants
Random Assignment It sin a procedure by which each person has the same likelihood to be assigned to any of the experimental groups (8, 5, hours of sleep etc) It can be done by pulling names out of a hat … but we use tables of random numbers When using manipulated independent variables we assign participants at random into the experimental groups. When using manipulated independent variables we assign participants at random into the experimental groups. Only by doing this we can say that it was the IV (sleep) that affected concentration and not any other variable
Non-Manipulated Independent Variables In many situations we are not able to randomly assign participants into groups The groups already exist For example: The impact of divorce of children’s academic performance In such cases we cannot determine cause-and-effect relationships In such cases we cannot determine cause-and-effect relationships The results are presented as correlations
Prediction vs. Hindsight In research we make predictions and then collect the data and interpret the findings Generating a hypothesis after the result are known (Hindsight”) is not regarded as scientific Replication is needed to examine participants populations different from the one used in the original research.
A Correlational Study A correlation is a procedure that tells us to what extent two variables vary together- that is- change in one variable is associated with change in another variable A correlation does not identify cause- and-effect relationships A correlation is described along its magnitude, from -1 to 1, and direction- positive, negative, or no correlation.
Statistical Significance A procedure that tells us the extent to which a result in an experiment reflects real differences among the groups, or chance fluctuations The Statistical Significance is expressed is terms of probability- e.g. result that is found to be significant at the 5% tells us that there is 95% probability that the result reflects true differences among the subjects, and not random factors
The Case Study This is the oldest form of study It involves studying one person in depth, often along time It is useful when there is a rare case that otherwise is difficult to study It is also useful to show the results of therapy It may also be subject to bias May be difficult to generalize to others
Psychological Assessment Personality tests need to possess psychometric qualities: Reliability: the extent to which the results of the test show consistency along time One form of assessing reliability is “test-retest” – administering the test to the same person at two different points in time Reliability also can be assessed as “internal consistency”- the extent to which the items of the test correlate with the overall score
Validity The extent to which the test measures what is claims to measure In developing anew test, validity needs to be demonstrated It is easier to demonstrate validity for some types of tests- a test (such as the SAT) to predict how well students will do in college needs to have predictive validity Those who score high on the SAT were found to have higher GPA in college
Validity (cont-d) Often is psychology we want to measure hypothetical constructs (such as intelligence), and look for Construct Validity The following methods are used to assess construct validity: Face Validity, Congruent Validity, Discriminant Validity, Behavioral Validation.
Face Validity We judge face validity by looking at the test items We say that “on the face of it” the tests measures the constructs For example, test to measure social anxiety will have items such as “Do you feel nervous interacting with others?” Some hypothetical constructs do not lend themselves to obvious questions.
Congruent Validity We compare the scores obtained on a new test to scores of the same people on an “old” test in the area We compare the scores obtained on a new test to scores of the same people on an “old” test in the area For example, when the new intelligence test (IQ) was introduced by David Wechsler in 1939, its validity was assessed by having people take the new test, and the old IQ test (Stanford-Binet) to see whether similar scores will be obtained The extent to which scores from the test correlate with other measures of the same construct
Discriminant Validity The extent to which a test score does NOT correlate with the scores of another test that is unrelated to it. Example: When a researcher designs a creativity test it is important to show that the test measures creativity instead of something different. Both tests are given to a group of people and the expectation is that correlations between the scores on the two tests in the people will be low.
Behavioral Validation The researcher gives the test, and then observes the behavior of interest in real life For example- give a test of Stage Fright and then observe actors just before they go on stage to perform. This is very relevant, as we often give a test in order to predict behavior.