No retraction, correction or concern found when last checked on Oct 24, 2026.

Analyses of ‘change scores’ do not estimate causal effects in observational data

Peter W G Tennant, Kellyn F Arnold, George T H Ellison and 1 more

International Journal of Epidemiology · 2021 · CC BY

Abstract

Background. In longitudinal data, it is common to create ‘change scores’ by subtracting measurements taken at baseline from those taken at follow-up, and then to analyse the resulting ‘change’ as the outcome variable. In observational data, this approach can produce misleading causal-effect estimates. The present article uses directed acyclic graphs (DAGs) and simple simulations to provide an accessible explanation for why change scores do not estimate causal effects in observational data.

Methods. Data were simulated to match three general scenarios in which the outcome variable at baseline was a (i) ‘competing exposure’ (i.e. a cause of the outcome that is neither caused by nor causes the exposure), (ii) confounder or (iii) mediator for the total causal effect of the exposure variable at baseline on the outcome variable at follow-up. Regression coefficients were compared between change-score analyses and the appropriate estimator(s) for the total and/or direct causal effect(s).

Results. Change-score analyses do not provide meaningful causal-effect estimates unless the baseline outcome variable is a ‘competing exposure’ for the effect of the exposure on the outcome at follow-up. Where the baseline outcome is a confounder or mediator, change-score analyses evaluate obscure estimands, which may diverge substantially in magnitude and direction from the total and direct causal effects.

Conclusion. Future observational studies that seek causal-effect estimates should avoid analysing change scores and adopt alternative analytical strategies.

Full text

Its licence does not allow us to show the full text. Read it at PubMed Central:

References 23

1.
Tu Y-K, Gilthorpe MS. Revisiting the relation between change and initial value: a review and evaluation. Stat Med 2007;26:443–57.
Not in our library10.1002/sim.2538
2.
Oldham PD. A note on the analysis of repeated measurements of the same subjects. J Chronic Dis 1962;15:969–77.
Not in our library10.1016/0021-9681(62)90116-9
3.
Senn S. Change from baseline and analysis of covariance revisited. Stat Med 2006;25:4334–44.
Not in our library10.1002/sim.2682
4.
Van Breukelen GJ. ANCOVA versus change from baseline had more power in randomized studies and more bias in nonrandomized studies. J Clin Epidemiol 2006;59:920–25.
Not in our library10.1016/j.jclinepi.2006.02.007
5.
Lord FM. A paradox in the interpretation of group comparisons. Psychol Bull 1967;68:304–05.
Not in our library10.1037/h0025105
6.
Glymour MM, Weuve J, Berkman LF, Kawachi I, Robins JM. When is baseline adjustment useful in analyses of change? An example with education and cognitive change. Am J Epidemiol 2005;162:267–78.
Not in our library10.1093/aje/kwi187
7.
Shahar E, Shahar DJ. Causal diagrams and change variables. J Eval Clin Pract 2012;18:143–48.
Not in our library10.1111/j.1365-2753.2010.01540.x
8.
Glymour MM, S Greenland. Causal diagrams. In: Rothman KJ, Greenland S, Lash TL (eds). Modern Epidemiology. Philadelphia: Lippincott Williams & Wilkins, 2008, pp. 183–212.
Not in our library
9.
Geiger D, Verma T, Pearl J. Identifying independence in Bayesian networks. Networks 1990;20:507–34.
Not in our library
10.
Bowman S. Radiocarbon Dating. Berkeley and Los Angeles: University of California Press/British Museum, 1990.
Not in our library
11.
VanderWeele TJ. Mediation analysis: a practitioner's guide. Annu Rev Public Health 2016;37:17–32.
Not in our library10.1146/annurev-publhealth-032315-021402
12.
Richiardi L, Bellocco R, Zugna D. Mediation analysis in epidemiology: methods, interpretation and bias. Int J Epidemiol 2013;42:1511–59.
Not in our library10.1093/ije/dyt127
13.
National Center for Health Statistics. National Health and Nutrition Examination Survey, 2009–2014 Data Files. Hyattsville: Centers for Disease Control and Prevention, 2016.
Not in our library
14.
Parametric versus non-parametric statistics in the analysis of randomized trials with non-normally distributed data · 2005
No notice
15.
Bias, precision and statistical power of analysis of covariance in the analysis of randomized trials with baseline imbalance: a simulation study · 2014
No notice
16.
Laird N. Further comparative analyses of pretest-posttest research designs. Am Stat 1983;37:329–30.
Not in our library
17.
O'Connell NS, Dai L, Jiang Y et al Methods for analysis of pre-post data in clinical research: a comparison of five common methods. J Biom Biostat 2017;8:1–8.
Not in our library10.4172/2155-6180.1000334
18.
Kim Y, Steiner PM. Gain scores revisited: a graphical models perspective. Sociol Methods Res 2019;004912411982615.
Not in our library
19.
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10.
Not in our library
20.
Gadd SC, Tennant PWG, Heppenstall AJ, Boehnke JR, Gilthorpe MS. Analysing trajectories of a longitudinal exposure: a causal perspective on common methods in lifecourse research. PLoS One 2019;14:e0225217.
Not in our library10.1371/journal.pone.0225217
21.
Beichler JE. The tie that binds: a fundamental unit of ‘change’ in space and time. In: Amoroso RL, Kauffman LH, Rowlands P (eds). The Physics of Reality: Space, Time, Matter, Cosmos. Singapore: World Scientific Publishing, 2013, pp. 19–28.
Not in our library
22.
Blastland M. The Hidden Half: How the World Conceals Its Secrets. London: Atlantic Books, 2019.
Not in our library
23.
Pearl J. Lord’s paradox revisited—(oh lord! kumbaya!). J Causal Inference 2016;4:2.
Not in our library

Reads

Read by people on this page

None yet.

Checked or read by AI through naulon

None yet.

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.

Are you an author?

Is this your paper?

Claim it to add your note and be paid when agents cite it. Free.