Field/experimental research , quantitative, attitudinal or descriptive, not archival.
Questionnaire versus Interview Administration.
Response types: Open ended and closed measures
Benefits to open ended measures:
Disadvantages:
Closed end measures: problems….don’t know or not appicable.
a) Rorschach inkblot test (open ended) b) Thematic Apperception Test (TAT) (wazzu open ended) c) Minnesota Multiphasic Personality Inventory (MMPI) (closed ended) 3. Rating Scales a) Item Numerical Scales stated numerical values or implied by order of response category assume equal intervals Criteria for construction a) levels must adequately sample the universe of content b) levels must be easily understood c) responses are in intended manner b) item forced-choice scales levels provided to the respondent push respondent by presenting equally favorable (unfavorable) values to them to choose. Attacks the Halo Effect c) Graphic Scales…line (segmented or not) answers. Bipolar adjectives at ends of line even number or odd number of segments depending upon need to push undecided respondents. 4. Rating Errors and their Control a) Halo effect is most prevalent when i) trait is not easily observed ii) is not clearly defined iii) involves relations : with other people iv) and is of some moral importance. v) Solution: forced choice scales b) Leniency bias: respondents rate someone with whom they are ego-involved. i) Solution: use forced choice scales with many adjectives on positive side. c) Central tendency bias: respondent hesitates to give extreme ratings. Ceiling and Floor effects. Eg. Three value labels may be: person is a ….louse, average guy, saint. i) Solution: expand the number of value levels on the scale from3 or 5 to 7 or 9 d) Logical error in rating: where respondent is rating a trait in someone else, their evaluation is their own not that of the subject. i) Solution: carefully define all terms used in rating scale. 5. Standardized Measures a) Semantic Differential: or How does a respondent perceive a thing. i) Assumes the “thing” has 3 underlying dimensions: overall evaluation, potency and activity. ii) Each underlying dimension is best captured by a bipolar scale: bad to good, weak to strong, slow to fast. iii) multiple items are used for each dimension b) The Likert Scale: Underlying attitude is complex, therefore needs many observations to capture subtle changes. i) Summated ratings methods of scale construction: ii) gather a large number of statements on the topic of interest iii) each statement is evaluated on a rating scale iv) statements are judged for fitness to be retained in the final scale by their consistency in measuring the intensity of evaluation : positive to positive statements, reversed to negative statements. v) Final scale has values ranging from the smallest possible from the sum of the items to the largest. c) Thurstone Scale: Similar to Likert Scale, Underlying dimension is complex, need many items to calibrate it. i) Take a large number of statements on a subject. Statement are evaluated on a bipolar scale. ii) Have judges place them into 11 piles each with equal judgmental distances between them. iii) A scale value is given to each statement. The median in the pile. iv) When administered, statements with a yes answer are used to compute a scale value for the respondent based upon the judge based values.