Recall bias is a systematic error that occurs when participants do not remember previous events or experiences accurately or omit details: the accuracy and volume of memories may be influenced by subsequent events and experiences. Recall bias is a problem in studies that have self-reporting, such as case-control study and retrospective cohort studies.
In case-control studies, researchers must be careful to ask each study participant in the same way so as not to influence their responses. Bias in recall can be greater when the study participant has poorer recall in general, and when the time interval being asked about is longer. Other issues that influence recall include age, education, socioeconomic status and how important the disease is to the patient. Added to this, undesirable habits such as smoking or eating unhealthy foods tend to be underreported, and are therefore subject to recall bias. Pre-existing beliefs may also impact on recall of previous events.
Parents of children diagnosed with cancer may be more likely to recall infections earlier in the child’s life than parents of children without cancer. This may lead to observing an entirely or partially untrue association between childhood infection and disease. Recall can be particularly problematic when the events of interest were a long time ago.
Recall bias can make observed associations bigger or smaller. For instance, when individuals recall lower rates of unhealthy food intake than the truth, associations will suggest lower levels increase risk. When people recall higher levels of fruit and vegetable intake (a desirable habit) than the truth, a protective association will suggest eating more is necessary to reduce their disease risk.
In the early 2000s, there was considerable publicity arising from a claim that the measles, mumps and rubella (MMR) vaccine was related to and possibly caused autism in children (the originating claim was subsequently found to be based on fraudulent data and the publication was withdrawn) (Andrews 2002). Researchers found that parents of autistic children diagnosed after the publicity tended to recall the start of autism as being soon after the MMR jab more often than parents of similar children who were diagnosed prior to the publicity.
A review of studies on emotionally arousing events found that victims of assault or war-exposure tended to amplify their memories of the events, while results from a vivid, enduring memory that was learned about through a surprising or shocking event (flashbulb memory) suggested memory for such emotional events is either stable or diminished over time.
A systematic review of bias in imaging studies found that the potential effects of observer recall bias is poorly researched in imaging and might introduce bias into radiological research.
Prospective cohort studies aim to avoid problems of recall of data by asking for information at the time that volunteers join the study. However, many also request historical information, which may be at risk of recall bias and researchers should be aware of this. Strategies that might reduce recall bias include careful selection of the research questions, choosing an appropriate data collection method, choosing people to study with new-onset disease or use a prospective design, which is the only way to ensure that recall bias is not present.
Case-control studies are useful for efficiency and in situations where the disease of interest is rare, but they are particularly at risk of recall bias. These studies need to make sure that the way questions are asked does not influence the participants’ answers.