230 likes | 342 Views
Chapter 5. Measuring Variables. From Hypothesis to Design. One of the first steps in designing an experiment is to operationally define the variables. IV – The causal variable There must be at least two different levels of the independent variable.
E N D
Chapter 5 Measuring Variables
From Hypothesis to Design • One of the first steps in designing an experiment is to operationally define the variables. • IV – The causal variable • There must be at least two different levels of the independent variable. • Presence (experimental treatment) vs. absence (control condition), or can have multiple levels • The differences are in AMOUNT or TYPE • Control Variables – Those that need to be held constant across levels of the IV • DV – The effect/outcome variable
Operational Definition • An explicit definition in terms that can be used to operate the research design • Operational definitions allow other researchers to replicate research exactly. • Operational definitions can be found by searching the literature. • For a dependent variable, the operational definition is a precise description of how it is measured.
Ways of asking questions • Fixed alternative questions • Questions where the respondent selects his or her answer from a set of specified responses. • Open-ended questions • Allows the respondent to provide his or her answer in his or her own words. • Often summarized with content analysis. • Rating Scales • Fixed alternative questions where the respondent indicates magnitude on a scale.
Ways of asking questions • Likert Scales • Statements in which the respondent is asked to indicate the degree he or she agrees or disagrees with the statement. • Semantic Differential • The respondent rates words on a series of bipolar scales. • Designed to measure the psychological distance between the connotative meanings of words.
Example • Alcohol and Aggression • Social Information Processing Studies • Free Response • What do we do with the answers? • Validity • Item Rating • How does this map onto behavior?
Social desirability and response set • Socially desirable responding – giving answers that respondents think are appropriate or that they think the researchers might want to hear • Response set – giving only moderate answers or always agreeing or disagreeing. • You can deal with these problems by: • Altering the wording on some items. • Stressing anonymity. • Including irrelevant items.
Before voting I thoroughly investigate the qualifications of all the candidates. • I never hesitate to go out of my way to help someone in trouble. • It is sometimes hard for me to go on with my work if I am not encouraged. • I have never intensely disliked someone. • On occasion I have had doubts about my ability to succeed in life. • I sometimes feel resentful when I don’t get my way. • I am always careful about my manner of dress. • My table manners at home are as good as when I eat out in a restaurant. • If I could get into a movie without paying and be sure I was not seen, I would probably do it. • On a few occasions, I have given up doing something because I thought too little of my ability.
Measuring observed behavior • Typical ways that psychologists measure observable behavior: • Accuracy – responses are either right or wrong. • Frequency – how often a behavior occurs in a specified period of time. • Latency – speed of onset. • Duration – how long the behavior lasts. • Amplitude – size of response. • Choice selection – frequency of choice between alternatives.
Example • Alcohol and Aggression Study • Amplitude • First Trial Shock Intensity • Average Shock Intensity • Playground Study • Behavioral Observation of Frequencies • Group Entry • Aggression • Anger and Cognition • RT for response to different word types
Levels of measurement • Quantitative variables • Differ in terms of amount; different values contain more or less of the variable. • Qualitative variables • Different values of these variables are different in quality. • Different levels of measurement reflect the degree of quantification of the measures.
Nominal • Differ in name (e.g.. gender, eye color). • Limited statistical manipulations. • Mode is best summary stat • Use frequencies, bar graphs • Chi-square is appropriate inferential stat
Ordinal • Vary in order of quantity (e.g.. first, second, and third place in a race). • Non-parametric statistical analyses. • Descriptive Statistics:frequency distributionmean, median, mode • Inferential Statistics:t test, analysis of variance (parametric)Wilcoxin rank-sum test, Kruskal-Wallace analysis of variance (nonparametric)
Interval • The intervals between the values of the variables are equal (e.g.. IQ scores). • Can use parametric statistical analyses. • Descriptive statistics:frequency distributionmean, median, mode • Inferential Statistics:t testanalysis of variance
Ratio • Like interval variables but with a true zero point (e.g.. temperature in Kelvin). • Can use the greatest variety of statistical analyses. • Descriptive statistics:frequency distributionmean, median, mode • Inferential Statistics:t testanalysis of variance
The importance of precision in measurement • In descriptive research, precise measurements can improve the accuracy of our description of a population. • When testing a research hypothesis, precision increases our chances of finding a statistically significant result. • Measures should be as precise as possible.
Reliability of measurement • Reliable measures are repeatable. • Test-retest reliability – determined by correlating the scores received by a group of participants who were tested at two different times. • Split-half reliability – determined by correlating half the items on a set of homogeneous questions with the other half. • Inter-rater reliability – measures the level of agreement between judges scoring behavior.
Validity of measurement • Face validity • Whether the measure seems to be a reasonable measure of the variable. • Construct validity • Whether the measure is measuring the underlying construct. • Determined by how well the measure of a variable fits into a theory.
Validity of measurement • Content validity • The degree that a measure assesses all the dimensions of the construct. • Criterion and predictive validity • If a measure of behavior has criterion validity then it should correlate with another measure of the same kind of behavior, a criterion measure.
Tests, scales, and inventories • Test – refers to many procedures used to measure a variable. • Scale – refers to a measure of a specific psychological characteristic. • Inventory – used to describe interests or personality.
Commercially available tests and inventories • Standardized tests of intelligence • Stanford-Binet Intelligence Scale (SBIS) • Wechsler Intelligence Scale for Children (WISC) • Wechsler Adult Intelligence Scale (WAIS) • Woodcock-Johnson Test of Cognitive Abilities • many more
Commercially available tests and inventories • Tests of personality • Standardized tests • MMPI • Millon Clinical Multiaxial Inventory (MCMI) • Sixteen Personality Factor Questionnaire (16PF) • California Psychological Inventory (CPI) • Projective tests of personality • Holtzman Inkblot Technique (HIT) • Rorschach Inkblot test • Thematic Apperception Test (TAT)
Commercially available tests and inventories • Other specialized tests • Bayley Scales of Infant Development (BSID) • Yesavage Geriatric Depression Scale (GDS) • Young Mania Rating Scale (YMRS) • many more