Numerous mental health questionnaires are available for researchers and practitioners to assess symptoms in a wide range of populations. For common mental health disorders such as depression or anxiety, a variety of questionnaires exist that differ by content or wording, who responds to the questions, how the questions are responded to, or the recall period (e.g., symptoms over the past week or month). These differences make integration between questionnaires difficult and can hinder estimations of prevalence, data pooling, data comparison, and more. Mental health research and practice can be significantly improved by effectively combining and comparing participant responses from different questionnaires.
Recently, researchers have tried to make questionnaires directly comparable through retrospective harmonization (Fortier et al., 2017; McElroy et al., 2020). This process involves modifying or recoding data from different questionnaires. One approach is to create subsets of harmonized items from various questionnaires that capture similar symptoms. In other words, although the wording of questions can differ across instruments, they often capture similar underlying experiences. For example, one questionnaire may ask whether a respondent “is miserable or tearful,” while another might ask if they are “often unhappy, down-hearted, or tearful.” Both statements capture a similar symptom (low mood), and recent psychometric research supports the equivalence of such items.
However, an aspect of questionnaire equivalence that has received less attention is the recall period. One study found that metric invariance was not supported when comparing “right now” and “in the past week” responses to a mood questionnaire, yet another study found that mean responses to the 7- or 30-day version of a psychological distress scale did not differ (Andrade & Rodríguez, 2018; Batterham et al., 2019). However, many earlier studies did not explicitly test for differential item functioning (DIF).
In this study described in a 2024 article published in Translational Issues in Psychological Science, Caitlyn Rawers, Marta Mangiarulo, Mark Shevlin, and Eoin McElroy focused on whether different recall periods to the Depression, Anxiety, and Stress Scale (DASS-21) resulted in DIF. The authors split participants into three groups and had them complete the DASS-21, with one group asked to recall their symptoms over the past week, another group asked over the past month, and the last group asked over the past 6 months. After examining each DASS-21 item, none of the questions displayed DIF, suggesting that participants responded similarly to the questions no matter the time frame for which they were considering their symptoms.
Many reasons exist for why people may respond similarly no matter the recall period specified, but cognitive biases or shortcuts (heuristics) have been shown to impact how people respond to mental health questionnaires, with people’s worst and most recent emotional states having a disproportionate effect on recall. Separate from the mechanism, this study’s findings suggest that questionnaires with different recall periods may be directly comparable, which can greatly improve efforts to compare and investigate mental health in different populations around the world.