Appraisal of Studies of High-Quality Sample Clauses

Appraisal of Studies of High-Quality. Quality scoring, particularly for observational research, is contestable (Greenland & X’Xxxxxx, 2001); however a decision was made to exclude studies with poor methodological quality that threatened the validity of findings. High-quality papers were defined as those that scored ≥50% on questions assessing selection bias. Selection bias was chosen as the primary indicator of high-quality because it examines biases in the representativeness of study samples, selection of controls in case-control studies, characteristics of participants and non-participants, and biases regarding differential loss-to-follow-up in cohort studies. Due to widespread variations in the terminologies used to measure domestic violence (see Chapter 1 section 1.5), assessing quality in relation to measurement biases was not considered to be a good indicator of the methodological strength of studies. A cut off score of 50% was selected as papers scoring below this figure were considered to possess too many biases, which would threaten the validity of findings (e.g. the results would not be representative of the population of interest). Reviewers compared scores and resolved any disagreements through consensus or with the aid of a third reviewer (LH) before calculating a final appraisal score. Scores for overall study quality and for questions relating to selection bias and measurement bias are reported for all studies (see Table 1).