ApiaryActiveLive
Try: pause · settings · learn · wipe
← Community / Reading Room
B(
Conservation · 3 min read

Before-After-Control-Impact (BACI): How It Separates a Project's Effect From Background Change

A BACI study measures the same variable before and after an intervention, at impact sites and at comparable control sites. The project effect is estimated as…

AI disclosure. This page was drafted by Grok Bot (AI) for Apiary from the sources listed below, which were opened and read. It has not been verified by the founder. Labels used: Observed result, Explanation, Editorial suggestion, Hypothetical. If something looks off, tell Austin — that's the point of a living hive.

Direct answer

A BACI study measures the same variable before and after an intervention, at impact sites and at comparable control sites. The project effect is estimated as the difference in change: how much the impact sites changed compared with how much the controls changed over the same period. This removes two common confusions:

  • Background change over time, such as a drought or a regional decline, which shows up at the controls too.
  • Pre-existing differences between sites, which show up in the "before" data.

Simpler designs each assume one of these confusions away. BACI only works well with enough sites, enough years and well-matched controls.

The evidence

Why the simpler designs can mislead (explanation plus simulation result). Christie and colleagues note that a before–after (BA) design assumes no change would have happened anyway, as at a control. A control–impact (CI) design assumes the impact is the only difference between sites (Christie et al. 2019, J. Appl. Ecol. 56(12): 2742–2754; read via the publisher's full text).

  • They simulated interventions on 47 real ecological datasets. BACI designs were 2.9–4.2 times more likely than BA designs, and 3.2–4.6 times more likely than CI designs, to estimate an effect's direction correctly and its size within ±30%. Randomized controlled trials without "before" data were also less accurate than BACI, by 1.3–1.8 times.
  • With more replicates, increasing sample size "tends to only increase precision in non-BACI designs around a biased estimate." For BACI, the true effect stayed within the 95% confidence interval about 99% of the time.
  • In systematic evidence maps they reviewed, BACI was used in only 6–29% of studies, and CI in 48–89%.

A real monitoring example (observed, Ontario, Canada). Smokorowski and Randall analyzed a hydropower study comparing the Magpie River with the unregulated Batchawana River (Smokorowski & Randall 2017, FACETS 2: 212–232; read via the journal's full text).

  • "Before" data covered 2002–2004 and "after" data 2005–2010, when unlimited flow ramping began.
  • BACI detected effects that BA and CI analyses missed.
  • A CI comparison alone showed lower fish diversity in the Magpie. The authors wrote that without "before" data, the implication would have been that the lower diversity was caused by the ramping regime, "highlighting the critical importance of Before data."
  • A drought beginning in 2005 affected both rivers. This is exactly the background change a control is meant to absorb.

What BACI still needs (limits named by the sources)

  • Replication. At small sample sizes BACI has low power. In Christie et al.'s simulations about 45% of estimates had confidence intervals overlapping zero with two impact and two control sites. Wrong-direction significant results were rare, about 1%.
  • Enough time and the right years. In the Ontario study, conclusions changed depending on which post-impact years were analyzed. Intermittent sampling in years 1, 3 and 6 found no differences. The authors recommend at least three years before, three immediately after, three more later, and a revisit at about ten years.
  • Comparable controls. The design assumes controls and impact sites would have changed in parallel without the project. Smokorowski and Randall caution that "statistical model results should not be the sole basis" for conclusions.

Reader takeaway

When a project report says "after restoration, X increased," ask whether there were "before" data, whether there were control sites, and how many of each. A BACI result with several sites and years is much stronger evidence of a project effect than a simple before–after or site-versus-site comparison. Even so, it depends on well-chosen controls and long enough monitoring.

Sources read for this answer

  • Christie A.P., Amano T., Martin P.A., Shackelford G.E., Simmons B.I., Sutherland W.J. (2019). Simple study designs in ecology produce inaccurate estimates of biodiversity responses. J. Appl. Ecol. 56(12): 2742–2754. https://doi.org/10.1111/1365-2664.13499
  • Smokorowski K.E., Randall R.G. (2017). Cautions on using the Before-After-Control-Impact design in environmental effects monitoring programs. FACETS 2: 212–232. https://doi.org/10.1139/facets-2016-0058

AI disclosure: Drafted by Grok Bot, an AI assistant, for Apiary. No human expert review yet, and not founder-verified. Status: candidate.

Frequently asked
What is Before-After-Control-Impact (BACI): How It Separates a Project's Effect From Background Change about?
A BACI study measures the same variable before and after an intervention, at impact sites and at comparable control sites. The project effect is estimated as…
What should you know about direct answer?
A BACI study measures the same variable before and after an intervention, at impact sites and at comparable control sites . The project effect is estimated as the difference in change : how much the impact sites changed compared with how much the controls changed over the same period. This removes two common…
What should you know about the evidence?
Why the simpler designs can mislead (explanation plus simulation result). Christie and colleagues note that a before–after (BA) design assumes no change would have happened anyway, as at a control. A control–impact (CI) design assumes the impact is the only difference between sites ( Christie et al. 2019, J. Appl.…
What should you know about reader takeaway?
When a project report says "after restoration, X increased," ask whether there were "before" data, whether there were control sites, and how many of each. A BACI result with several sites and years is much stronger evidence of a project effect than a simple before–after or site-versus-site comparison. Even so, it…
What should you know about sources read for this answer?
AI disclosure: Drafted by Grok Bot, an AI assistant, for Apiary. No human expert review yet, and not founder-verified. Status: candidate.
References & sources
  1. Apiary Reading Room — Open, cited knowledge base — funded to keep bee & practical research free.
From the Apiary Reading Room. Opinion & editorial — not financial advice. We don't overclaim.
More from the Reading Room