Running Head: PSYCHOMETRIC ANALYSE OF THE GARED TEST
Psychometric Analyse Of The Girls At Risk Of Eating
Disorders (GARED)
Cedric Badjeck, 2646177
Measurement Theory and Assessment 2
Vrij Universiteit Amsterdam
, PSYCHOMETRIC ANALYSE OF THE GARED
Psychometric Analyse Of The GARED
Eating disorders during childhood predict a risk to health and wellbeing later in
adolescence. Therefore developing tests that are specially designed to identifying factors that
might place young girls at risk of developing eating disorders is paramount. One of such tests
to detect factors that might place young girls at risk of developing eating disorders is the
Girls At Risk Of Eating Disorders test (GARED). One aim of the GARED was to develop a
screening instrument that is administered as a online test.
The GARED test consists of 15 items with an ordinal 5-point scale response format. In
addition to the test items, an external measure of risk was obtained, based on a 20-minute
clinical interview, which was conducted individually by trained MA-level psychologists at the
girls’ school. The risk variable is dichotomous, with values 0 (not at risk) and 1 (at risk).
Because the GARED is an online test, it is important that the test be short. The first
aim of this study is to investigate whether a shortened version of the test can be used as a
screening tool for the tendency of developing an eating disorder. By conducting a factor
analysis and looking at the items with low communalities, we removed 5 items that were least
useful to predict the tendency of developing eating disorders. Communalities show the
amount of variance in the item that is explained by a common factor.
The second aim of this study is to determine the test’s dimensionality of the remaining
10 items. A test is unidimensional when all the items are all dependent on the same variable,
and multidimensional when the responses to the items depend not only on one variable but
also on another latent variable.
The third aim of this study is to determine whether the GARED test scores are
predictive of the risk of developing an eating disorder. We performed a logistic regression
analysis in SPSS. The results reveal the probability of being at risk of developing an eating
disorder when a certain GARED test score is obtained.
Samples and variables
A sample of 908 girls completed the GARED test online. Their age ranged from 13 to
17 years old (M=14.98, SD=1.45). The 15 items of the GARED test are shown in Table 1.
Table 1:
The 15 GARED Test items.
Item 1 I dislike it when people tell me I should eat more
Item 2 I usually know exactly what I will eat during the day
Item 3 I dislike it when people tell me I should eat more
Item 4 My family and friends say I look normal, but I feel fat
Psychometric Analyse Of The Girls At Risk Of Eating
Disorders (GARED)
Cedric Badjeck, 2646177
Measurement Theory and Assessment 2
Vrij Universiteit Amsterdam
, PSYCHOMETRIC ANALYSE OF THE GARED
Psychometric Analyse Of The GARED
Eating disorders during childhood predict a risk to health and wellbeing later in
adolescence. Therefore developing tests that are specially designed to identifying factors that
might place young girls at risk of developing eating disorders is paramount. One of such tests
to detect factors that might place young girls at risk of developing eating disorders is the
Girls At Risk Of Eating Disorders test (GARED). One aim of the GARED was to develop a
screening instrument that is administered as a online test.
The GARED test consists of 15 items with an ordinal 5-point scale response format. In
addition to the test items, an external measure of risk was obtained, based on a 20-minute
clinical interview, which was conducted individually by trained MA-level psychologists at the
girls’ school. The risk variable is dichotomous, with values 0 (not at risk) and 1 (at risk).
Because the GARED is an online test, it is important that the test be short. The first
aim of this study is to investigate whether a shortened version of the test can be used as a
screening tool for the tendency of developing an eating disorder. By conducting a factor
analysis and looking at the items with low communalities, we removed 5 items that were least
useful to predict the tendency of developing eating disorders. Communalities show the
amount of variance in the item that is explained by a common factor.
The second aim of this study is to determine the test’s dimensionality of the remaining
10 items. A test is unidimensional when all the items are all dependent on the same variable,
and multidimensional when the responses to the items depend not only on one variable but
also on another latent variable.
The third aim of this study is to determine whether the GARED test scores are
predictive of the risk of developing an eating disorder. We performed a logistic regression
analysis in SPSS. The results reveal the probability of being at risk of developing an eating
disorder when a certain GARED test score is obtained.
Samples and variables
A sample of 908 girls completed the GARED test online. Their age ranged from 13 to
17 years old (M=14.98, SD=1.45). The 15 items of the GARED test are shown in Table 1.
Table 1:
The 15 GARED Test items.
Item 1 I dislike it when people tell me I should eat more
Item 2 I usually know exactly what I will eat during the day
Item 3 I dislike it when people tell me I should eat more
Item 4 My family and friends say I look normal, but I feel fat
