Control groups

When researchers are merely interested in the linguistic outcomes of any given language intervention, an active control group completing a different intervention (e.g., playing guitar) may not be necessary. Instead, one could opt for a within-group development design, or compare the language development in the intervention group to a passive control group that did not complete any intervention. An exception to this is when researchers are intent on comparing the efficacy of different language learning pedagogies on language development, in which case participants receive one of two (or more) teaching methods, respectively (see, van der Ploeg, Lowie, and Keijzer 2023, although they also included cognitive and socio-affective measures).

To acknowledge the fact that aging is a multidimensional phenomenon (e.g., Christopher 2014), researchers may opt for an integrative approach, where both cognitive, socio-affective, and linguistic outcome measures are collected. This approach reflects that learning a new skill (i.e., the language taught in the course) may also have a positive effect on more domain-general cognitive functioning and well-being. For example, if a learner does not improve their proficiency in the language that was taught, but their cognitive flexibility does improve, the learning experience itself may still be a potentially useful tool to stimulate healthy aging. If an additional goal of a given LLLL study is to disentangle the effects of doing something cognitively stimulating from the purported unique effects of learning a language, it is advised to include an active control group. Control interventions in the past have included: art workshops (Nijmeijer et al. 2021), learning to play the guitar (Nijmeijer et al. 2021), following lectures (Brouwer et al., in preparation), and playing (video)games (P. C. M. Wong et al. 2019; Kliesch et al. 2022). The choice of control intervention should ideally be based on evidence of induced cognitive advantages in the older adult lifespan, as is the case for example with musical training (Marie et al. 2023).

Furthermore, potential participants may strongly prefer one of the interventions offered by the study in case of there being multiple study arms. In order to see if such a preference exists in the intended study sample, one could consult the target population prior to recruitment to see if they would like to participate in the control intervention to begin with. This can be done, for example, by using focus groups. From practice, we know that not every control intervention is enjoyed equally. For example, the music intervention in one of our studies (Brouwer et al., in preparation), a guitar course, was not a popular option; several potential participants refused to participate knowing that group-allocation would be random, and some participants dropped out before the start of the course when they heard that they had been assigned to the guitar course. Likewise, during the guitar course, we noticed that several participants dropped out due to physical discomfort they experienced in their fingers and hands.


Further reading