The Developmental Object Familiarity Inventory (DOFI)

Altmann, E. C. and van Renswoude, Daan and Raijmakers, Maartje (2020) The Developmental Object Familiarity Inventory (DOFI). In: International Conference of Infant Studies 2020, 2020-07-06 - 2020-07-09, online.

Abstract

Numerous studies have used object familiarity as an independent variable without ever actually defining the construct. Instead, it has been used as a measure of exposure - more exposure of a stimulus making it more familiar. Yet, we argue, that being familiar with an object is more than simple exposure: different levels of experience with and knowledge about an object should be integrated into the construct. Thus, we created the Developmental Object Familiarity Inventory (DOFI) as a new parent report measure to evaluate object familiarity on a six-point scale. The scale points are: (1)"My child has never seen this object before", (2)"My child has paid attention to this object before", (3)"My child has shown interest to explore the object at least once", (4)"My child has some knowledge about the object's use", (5)"My child can indicate where the object is when asked for it", and (6)"My child has and uses a consistent word for the object". This way, real-life familiarity about real-life objects can be measured, as well as, other research questions related to object familiarity (e.g., van Renswoude et al., 2019). Items were 76 objects covering six object categories following the CDI's organization (Fenson et al., 1993): 'vehicles' (9 items), 'toys' (9 items), 'food' (11 items), 'clothing' (14 items), 'household' (18 items), and 'furniture' (15 items) based on the Dutch CDI (N-CDI; Zink & Lejaeger, 2002). In three studies the new measure's reliability and validity was investigated by collecting data about infants between the age of 6 and 24 months. A first study (N = 10 infant reports, M = 15,37; SD = 3,98) shows that the six scale points indeed follow a strict order for all parents for all items. A second study (N = 28 complete infant reports, M = 13,5; SD = 5,5) was aimed at testing the internal structure of the questionnaire (Messick, 1995; Downing, 2003). The survey had great internal consistency (Cronbach's alpha is .991) and a good interitem correlation (d = .63). The scale's ordinality was to be analyzed with a graded response model (GRM, Samejima, 2011; Figure 1). When including age and gender as covariates in the model, age has a significant, positive effect on the latent trait (β = 1.17, t = 2.67, p < .05). Twelve participants responded to the test-retest survey. On average, the interval between measurements was 23 days (min = 9; max = 57; SD = 12,64). With a correlation of r = .86 (p < .001) the test-retest reliability was good. In the third, pilot study, an eye-tracking validation (N = 6 infants between the ages of 7,3 and 22 months, M = 14,4; SD = 4,52), the familiarity scores obtained in the DOFI were attempted to be related to preferential looking. We gathered important information to consider when designing future studies implementing the DOFI (in progress). In conclusion, the DOFI shows great potential to be used in object familiarity related research and might revolutionize the field towards a more natural and generalizable manner of research.