HANDOUT 7.3

THE CAUSES OF NULL RESULTS IN
CORRELATIONAL RESEARCH ARE SIMILAR TO THE
CAUSES FOR NULL RESULTS IN EXPERIMENTAL
RESEARCH


PROBLEM IN CORRELATIONAL
RESEARCH
PARALLEL IN EXPERIMENTAL RESEARCH
NOT ENOUGH PARTICIPANTS NOT ENOUGH PARTICIPANTS
INSENSITIVE MEASURES INSENSITIVE MEASURES
NON-LINEAR RELATIONSHIPS NON-LINEAR RELATIONSHIPS -
ESPECIALLY A PROBLEM WHEN
INVESTIGATOR DOING SIMPLE
EXPERIMENT CHOOSES A LOW LEVEL
OF THE INDEPENDENT VARIABLE AND
A HIGH LEVEL OFTHE INDEPENDENT
VARIABLE, AND THE RELATIONSHIP
IS NOT LINEAR.
RESTRICTION OF RANGE IN THE
1. PREDICTOR VARIABLE
(NOT CHOOSING PARTICIPANTS WHO WERE
DIFFERENT ENOUGH ON THE
VARIABLE) OR
2. CRITERION VARIABLE,
DUE TO CEILING EFFECTS, FLOOR
EFFECTS, INSENSITIVE MEASURES,
OR NOT CHOOSING PARTICIPANTS WHO
WERE DIFFERENT ENOUGH ON THAT
VARIABLE.
NOT HAVING ENOUGH VARIABILITY
IN THE
1. INDEPENDENT VARIABLE
(NOT ASSIGNING PARTICIPANTS TO
LEVELS THAT WERE DIFFERENT
ENOUGH ) OR
2. DEPENDENT VARIABLE,
OFTEN DUE TO CEILING EFFECTS,
FLOOR EFFECTS, OR AN
INSENSITIVE MEASURE.


Back to Chapter 7 Main Menu