Back to Psyc 1B | Back to Jeanie's home

Chapter 3

Steps of the Scientific Method

1.     Develop a research question(s), usually based on a thorough review of a theory or an area of interest.

2.     Formulate an hypothesis.

3.     Design a study to test the hypothesis.

4.     Conduct a study and collect data.

5.     Evaluate the data in terms of the prediction and hypothesis.

6.    

Variables

Variable: any event, situation, behavior, or individual characteristic that varies (can have more than one value).

¨.

¨.

Theories

n    Good theories are parsimonious and comprehensive. 

n    .

n    May be replaced by newer theories that provide better explanations of behavior.

n   

n    They help organize and unify different observations related to the behavior

n   

Constructs

n   Hypothetical attributes or mechanisms that help explain and predict behavior in a theory.

n  

Criteria for Operational Definitions

Any mental construct we wish to study must be defined so that it is:

1.        

2.        

3.        

Today you are going to practice doing psychological research

First Problem…

n  

¨

Initial Stages of Research

n.

n.

nresearch hypothesis: a tentative statement (a prediction or an educated guess) about the relationship between 2 or more variables.

¨

Next problem…

 

n  

n   concepts must be operationally defined.

¨

How can we measure anxiety?

n   What observable behaviors might demonstrate that a person is anxious?

How can we measure anxiety?

n   What observable behaviors might demonstrate that a person is anxious?

Can we measure anxiety in another way?

Multiple Measures

Reliability and Validity

n   Reliability (consistency):

¨ A reliable measure

n  .

n   Validity:    

¨ A valid measure

n .

Reliability

n   There are two components of a score:

¨ A true score

¨ Measurement error

Measuring Reliability

n     There are 4 main ways to measure the reliability of a measure:

1. Test-Retest reliability

n     You collect 2 scores from a large random sample of subjects at 2 different points in time.

n     Problem

¨   .

Measuring Reliability

2. .

Measuring Reliability

n          Internal consistency

n          .

n          You can randomly divide items on a test into two groups (split-half reliability).

Measuring Reliability

n          Interrater reliability

n          . 

Validity

n   

n    Construct validity

n    A measurement tool is valid if it really measures the construct it is intended to measure rather than some other characteristic.
 

Face Validity

n         How well does the test appear to measure what it was designed to measure based on its contents?

n        

n         Based on the subjective evaluations of the assessor.

n         .

Convergent Validity

n          The scores of a measure correlate with the scores on another established measure of the same construct.

n         

Divergent (discriminant) Validity

n        The scores of a measure DO NOT correlate with scores on another established measure of a different construct.

n       

Predictive Validity

n   The measure allows you to predict future behavior.

n   .

Concurrent Validity

n   Two or more groups of people differ on the measure in expected ways.

n   .

Measurement Scales

Nominal Scales

n    involve simply categorizing the stimulus to be measured into one of a number of discrete qualitative categories.           

n   

n   

Ordinal Scales

n    Ordinal scales add the arithmetic relationships "greater than" and "less than" to the measurement process. 

n    examples:

Interval Scales

n    involve specifying how far apart 2 stimuli are on a given dimension. 

n     

n    Equal units of measurement.

n    example:

Ratio Scales

n    A ratio scale has all the properties of an interval scale with the important additional quality of an absolute zero point. 

n   

n    In psychology, we commonly do not see absence in behavior except when measuring time.

Class Activity

Measurement

n   Sensitivity

n   Range Effects

¨

¨ Floor effect

Reactivity

n    Reactivity in measurement occurs when the very act of measuring a variable changes an individual’s behavior.

v  . 

vUsing unobtrusive measures reduces the   likelihood of measurement reactivity.

n    Demand Characteristics

Experimenter bias

 

n   Researchers influence the results

¨

Handling reactivity & experimenter bias

n  

n