Experimental Psychology
Ad 0:
Digital Ocean
Providing developers and businesses with a reliable, easy-to-use cloud computing platform of virtual servers (Droplets), object storage ( Spaces), and more.
2005-02-21 17:48:25 (UTC)

Hill 2-21

Evaluating Measures: Validity

-Is it an "appropriate" measure of the construct?
-Operational measures must be valid or directly refer to
what they are supposed to
-Types of validity
-Face validity- self-evident, obviously related to
-E.g., intelligence test-has problems to solve;bike test-
ride b/w 2 lines
-content validity - does "content" of measure redlect
"content" of what is being measured?
-measure of questions are relevant
-Predictive/criterion validity - do procedures yeild
info that would predict behavior in another situation;
or is it related to another simular measure
-can it classify hi vs. low groups?(e.g.,IQ to school)
-Construct validity - most important does instrument
measure the construct? & is construct being measured

Reliability & Validity
Together at last

-Both are important
-Both are correlational assesments
-Reliability limits validity
-Measures have to be reliable to be valid
-BUT, measures can be reliable and not valid
-E.G., Phrenology-measuring skulls

Scales of measurment
-Differnet ways of assigning numbers to variables
-Type of measurment scale dictates the appropriate
statistical analysis
-E.g., catgories
-Ordianal-#= rank;relative position
-E.g., class rankings, best love song of all time vs best
song of all time
-Interval-rank order w/meaningful, equel intervals, no
zero (size)
-e.g., personality tests, attitude inventories, IQ tests
-Ratio -ordered scales, equel intervals, and a true zero
-e.g., height, weight, time# errors made in task

The Definition of an Experiment

-Goal: to explain an outcome
-The most systematic form of research
-Only type to assess causality
-"IF" - manipulate (antecedent) treatment conditions (IV)
-Control extraneous variables (EV)
-"THEN" - Measure of behavior (DV) under each condition
to see if treatment had an effect

-If experiment is carefully designed , conducted and
analized then able to infer cause (i.e., internal
validity) and generalize results to population (i.e.,
external validity)

Independant Variables (IV)

-Factor of interest to researcher that is manipulated
(influenced/managed) by E
-Must have at least 2 levels (treatment conditions)
-Types of IV's (antecedints)
-situational -aspect of environmental conditions not
just "conditions"
-E.g.,# of bystandards (0,3,6)
-Task - what participants asked to do, different problems
-E.g.,maze types, memory stratagies
-Instructional - asked to act differently on same task
-E.g.,use a different encoding stratagy
-Organismic(i.e., ex post facto, naural group.
nonmanipulated. subject)
-already determaned, aspect of Ss cant manipulate