Quality & Quantity 32: 63–75, 1998.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.
Short Time Series Analysis:
C Statistic vs Edgington Model
JAUME ARNAU & ROSER BONO
University of Barcelona
Abstract. Young’s C statistic (1941) makes it possible to compare the randomization of a set of
sequentially organized data and constitutes an alternative of appropriate analysis in short time series
designs. On the other hand, models based on the randomization of stimuli are also very important
within the behavioral content applied. For this reason, a comparison is established between the C
statistic and the Edgington model. The data analyzed in the comparative study have been obtained
from graphs in studies published in behavioral journals. According to the results obtained, it is
concluded that the Edgington model in experimental designs AB involves many measurements while
the C statistic requires fewer observations to reach the conventional signiﬁcance level.
Key words: short time series, C statistic, Edgington models.
It is easy to verify from journals having an applied character that most of the
studies done with single-subject designs generally use few observations per phase.
Considering this fact, it has been demonstrated that the application of the inter-
rupted time-series analysis (ITSA) is not recommendable because the power of
this procedure decreases as the amount of data is reduced (Harrop & Velicer, 1985;
Tryon, 1982). Thus, over the last 25 years, a series of alternative statistical models
has been developed that makes it possible to infer the effect of the treatment in
short time series designs (Algina & Swaminathan, 1977; Edgington, 1975, 1980,
1992; Gorsuch, 1983; Huitema & McKean, 1991; Kazdin, 1976; Levin, Marascuilo
& Hubert, 1978; Simonton, 1977; Tryon, 1982, 1984). This study establishes a
comparison between the C statistic (DeCarlo & Tryon, 1993; Tryon, 1982) and
the Edgington model applied to AB designs (Edgington, 1980, Onghena, 1992). In
this way, two nonparametric type tests based on the principles of randomization
According to the logic of the Edgington model (1980), the treatment point is
randomly determined. The random decision of the intervention interval deﬁnes the
limit between the amount of times that a record is taken under the condition of
no treatment (phase A or baseline phase) and the amount of observations under
the action of treatment (phase B or experimental phase). On the other hand, it
can be expected that the experimenter is not interested in using the observation
occasions located on the extremes of the series as an intervention point. In this