Running head: A PSYCHOMETRIC ANALYSIS OF THE GARED TEST 1
A Psychometric Analysis of the GARED Test
Mandy Roosendaal, 2663488
Measurement Theory & Assessment II
Diana Lucía Gross, Group 9
Word count: 1551
20-12-2020
, A PSYCHOMETRIC ANALYSIS OF THE GARED TEST 2
A Psychometric Analysis of the GARED Test
The Girls At Risk of Eating Disorders (GARED) study aimed to develop an online
screening instrument to help identify girls who are at risk of developing an eating disorder. From
this study, the GARED test resulted. This test contains 15 items with a 5-point response scale in
which participants indicate whether the item applies to them. For this report, a psychometric
analysis of the GARED test is carried out in the Statistical Package for the Social Sciences
(SPSS).
An important aspect of online testing is that the test must be short. With 15 items, the test
is currently too long. Therefore, the first aim of this analysis is to remove five items from the
test. To determine which five items of the test are least useful in predicting the risk of developing
an eating disorder, and are thus best suited for removal, a factor analysis is carried out in SPSS
which reveals the communalities of all items. Communalities indicate the amount of variance in
the item that is explained by the risk of developing an eating disorder. The items with the lowest
communalities are thus least related to the risk of developing an eating disorder and, therefore,
the five items with the lowest communalities are removed from the test.
The second aim of this psychometric analysis is to determine the test’s dimensionality. By
means of a factor analysis of the ten remaining items, it will be determined whether the items are
all dependent on the same latent variable. If this is the case, the test is unidimensional, and the
responses to the items all depend on the risk of developing an eating disorder. If not, the test is
multidimensional, and the responses to the items depend not only on the risk of developing an
eating disorder but also on another latent variable.
The third and final aim of this analysis is to determine whether the GARED test scores
are predictive of the risk of developing an eating disorder. A logistic regression analysis is carried
A Psychometric Analysis of the GARED Test
Mandy Roosendaal, 2663488
Measurement Theory & Assessment II
Diana Lucía Gross, Group 9
Word count: 1551
20-12-2020
, A PSYCHOMETRIC ANALYSIS OF THE GARED TEST 2
A Psychometric Analysis of the GARED Test
The Girls At Risk of Eating Disorders (GARED) study aimed to develop an online
screening instrument to help identify girls who are at risk of developing an eating disorder. From
this study, the GARED test resulted. This test contains 15 items with a 5-point response scale in
which participants indicate whether the item applies to them. For this report, a psychometric
analysis of the GARED test is carried out in the Statistical Package for the Social Sciences
(SPSS).
An important aspect of online testing is that the test must be short. With 15 items, the test
is currently too long. Therefore, the first aim of this analysis is to remove five items from the
test. To determine which five items of the test are least useful in predicting the risk of developing
an eating disorder, and are thus best suited for removal, a factor analysis is carried out in SPSS
which reveals the communalities of all items. Communalities indicate the amount of variance in
the item that is explained by the risk of developing an eating disorder. The items with the lowest
communalities are thus least related to the risk of developing an eating disorder and, therefore,
the five items with the lowest communalities are removed from the test.
The second aim of this psychometric analysis is to determine the test’s dimensionality. By
means of a factor analysis of the ten remaining items, it will be determined whether the items are
all dependent on the same latent variable. If this is the case, the test is unidimensional, and the
responses to the items all depend on the risk of developing an eating disorder. If not, the test is
multidimensional, and the responses to the items depend not only on the risk of developing an
eating disorder but also on another latent variable.
The third and final aim of this analysis is to determine whether the GARED test scores
are predictive of the risk of developing an eating disorder. A logistic regression analysis is carried
