Peer Review #1 of "Regression assumptions in clinical psychology research practice—a systematic review of common misconceptions (v0.1)" [peer_review]

M Williams
2017 unpublished
Misconceptions about the assumptions behind the standard linear regression model are widespread and dangerous. These lead to using linear regression when inappropriate, and to employing alternative procedures with less statistical power when unnecessary. Our systematic literature review investigated employment and reporting of assumption checks in twelve clinical psychology journals. Findings indicate that normality of the variables themselves, rather than of the errors, was wrongfully held for
more » ... a necessary assumption in 4% of papers that use regression. Furthermore, 92% of all papers using linear regression were unclear about their assumption checks, violating APA-recommendations. This paper appeals for a heightened awareness for and increased transparency in the reporting of statistical assumption checking. PeerJ reviewing PDF | Manuscript to be reviewed 1 2 3 Regression Assumptions in Clinical Psychology Research Practice -A systematic review of 4 common misconceptions 5 6 Abstract 7 Misconceptions about the assumptions behind the standard linear regression model are 8 widespread and dangerous. These lead to using linear regression when inappropriate, and to 9 employing alternative procedures with less statistical power when unnecessary. Our systematic 10 literature review investigated employment and reporting of assumption checks in twelve clinical 11 psychology journals. Findings indicate that normality of the variables themselves, rather than of 12 the errors, was wrongfully held for a necessary assumption in 4% of papers that use regression. 13 Furthermore, 92% of all papers using linear regression were unclear about their assumption 14 checks, violating APA-recommendations. This paper appeals for a heightened awareness for and 15 increased transparency in the reporting of statistical assumption checking. 16 17 PeerJ reviewing PDF | Manuscript to be reviewed 18 Regression Assumptions in Research Practice -A systematic review of common misconceptions 19 20 One of the most frequently employed models to express the influence of several 21 predictors on a continuous outcome variable is the linear regression model: 22 23 This equation predicts the value of a case Y i with values X ji on the independent variables X j (j = 24 1, ..., p). The standard regression model takes X j to be measured without error ( cf. Montgomery, 25 Peck & Vining, 2012, p.71). The various β j slopes are each a measure of association between the 26 respective independent variable X j and the dependent variable Y. The error for the given Y i , the 27 difference between the observed value and value predicted by the population regression model, is 28 denoted by ε i and is supposed to be unrelated to the values of X p . Here, β 0 denotes the intercept, 29 the expected Y value when all predictors are equal to zero. The model includes p predictor 30 variables. In case p = 1, the model is denoted as the simple linear regression model. 31 The standard linear regression model is based on four assumptions. These postulate the 32 properties that the variables should have in the population. The regression model only provides 33 proper inference if the assumptions hold true (although the model is robust to mild violations of 34 these assumptions).
doi:10.7287/peerj.3323v0.1/reviews/1 fatcat:t6khzdv7dze5tbc7xnhzrqhawu