|
Some individuals may be interested in the
statistical details behind these questionnaires.
The following information, while a bit detailed,
may be helpful to individuals who must document
the strength and appropriateness of these
outcome questionnaires. The following points
highlight the construction and testing of
the questionnaires:
| A
classical test theory foundation
(e.g., DeVellis, 2003; Nunnally & Bernstein,
1994; Suen, 1990; Cronbach, 1990)
was used in the four phases of this
study: (1) identification of target
outcomes, (2) operationalization
of those outcomes, (3) data collection,
and (4) data analysis and interpretation.
The questionnaires went through rigorous
psychometric testing during the pilot
testing phase (summer 2006) through
the participation of eleven camps
and a total of 791 campers. Psychometric
evaluation included examination of
the internal structure (reliability,
item-to-total correlations, inter-item
correlations) of the individual instruments
as well as cross-structure analysis
(intercorrelations among scale scores
and correlations between scale scores
and age) |
| |
| Analysis
of data supported the position that
the Camper Learning survey can appropriately
used as a unidimensional measure of
generalized camp learning. The alpha
reliability coefficient was .85. Item-to-total
correlations ranged from .29 to .61.
Thirteen of the 14 coefficients exceeded
.40, and eight of those exceeded .50.
Random effects analysis of variance
revealed no significant “interviewer” variance. |
| |
| Data from
the Basic Camp Outcomes Questionnaires
suggested that this format worked well.
All reliability coefficients exceeded
.90, and all item-to-total correlations
were greater than .50. The final versions
of each of these seven scales include
6-14 items and are scored on five point
rating scales. |
| |
| The Detailed
Camp Outcomes Questionnaires format
also worked very well. The change versions
produced alpha reliability coefficients
ranging from .87 to .93, and six of
the seven alpha coefficients for the
perceived status measures were above
.80. Criterion related evidence of
validly based on the matrix of correlations
among the scale scores supported the
use of these scales for their intended
purpose. The final versions of each
of these seven scales (status and change)
include 6-13 items and are scored on
six point rating scales. |
| |
| Return to previous
page |
Top of page |
 |
|
