Show Us the Data

Authoritative statements in ACI documents are often unaccompanied by any reference to published data that validate the statements. That’s a problem.

Reinforcing steel is so congested in the shear wall boundary element (right side of photo) that concrete must be placed in the less heavily reinforced shear wall section (left side of photo) and vibrated at that point to move it laterally into the boundary element.
Reinforcing steel is so congested in the shear wall boundary element (right side of photo) that concrete must be placed in the less heavily reinforced shear wall section (left side of photo) and vibrated at that point to move it laterally into the boundary element.

Many American Concrete Institute (ACI) documents, especially guides and reports, contain statements unaccompanied by any cited data that supports them. Some of these are remnants of language used in earlier versions of the documents, which may or may not have cited data or explained the rationale for the statement. In the consensus process for developing such documents, anecdotal evidence is almost certainly responsible for some of the unsupported statements. But we know that any beneficial use of anecdotal evidence in creating guides for construction can be limited by the number of uncontrolled — and unreported — variables present on every jobsite.

Construction tolerances based on what?

Section 9.4 of ACI 522R-10, “Report on Pervious Concrete,” includes the following statement:

“Tolerance for thickness reported as the average of three cores of each lot should not be more than 1/4 in. (6 mm) less than the specified thickness, with no single core exceeding 1/2 in. (13 mm) less than the specified thickness, nor should the average compacted thickness be more than 1.5 in. (40 mm) more than the specified thickness.”

There is no citation or reference to data that shows this to be a reasonable tolerance. But there is a wealth of data showing that this is an unreasonable tolerance for concrete floor slabs, which are usually placed on a smoother base course than the open-graded base courses often used under pervious pavements.1

So how was the tolerance set: Arbitrarily? Using anecdotal evidence from a committee member or committee members? Remember the often used statement that two anecdotes don’t comprise a data set.

What causes layer lines?

Section 3.8 of ACI 309.2R-98, “Identification and Control of Visible Effects of Consolidation on Formed Concrete Surfaces,” defines layer lines and causes as follows:

“Layer lines are dark horizontal lines on formed surfaces that indicate the boundary between concrete placements. Layer lines are caused by premature stiffening or insufficient consolidation of the previous layer of concrete due to lack of penetration of the vibrator into that layer, or the use of a mortar bonding layer between placements.”

Again, no citation shows that the list of causes is complete or indicates the conditions that lead to conclusions about causes. As stated by Steven Novella in his article in Science-Based Medicine:

“The primary weakness of anecdotes as evidence is that they are not controlled. This opens them up to many hidden variables that could potentially affect the results. We therefore cannot make any reliable assumptions about which variable (for example a specific treatment) was responsible for any apparent improvement.”2

Are there hidden variables that affect the formation of layer lines? We think so.

Effects of form release agents on rebar

Section 5.3 of ACI 304R-00, “Guide for Measuring, Mixing, Transporting, and Placing Concrete,” states the following:

“At the time of concrete placement, reinforcing steel and embedded items should be clean and free from mud, oil, and other materials that can adversely affect the steel’s bonding capacity.”

This statement is based on a nearly identical requirement that has appeared for many years in ACI 318, “Building Code Requirements for Structural Concrete.” The Code uses the phrase “…that decrease bond,” but in the ACI 318 Commentary, no publications with data are cited to show that oil, in particular, decreases bond. Research published in 1968 showed that a normal amount of rust on reinforcing steel increased bond. This research is cited in the ACI 318 Commentary. Research on the effect of form release agents on bond showed that there was no adverse effect.3 However, contractors are often required to remove form release agents from rebar, usually increasing labor costs and slowing progress.

Moving concrete laterally

Section 5.4.1 of ACI 304R-00 states that concrete should be deposited at or near its final position because it tends to segregate when it has to be flowed laterally into place. Section in the 2007 ACI Manual of Concrete Inspection includes an even stronger statement:

“The operator should not use the vibrator to cause concrete to flow from one location to another, because such a practice usually causes segregation.”

We could find no published evidence or data showing that flowing concrete laterally usually causes segregation. In fact, the ACI Committee E-703 publication Cast-in-Place Walls4 states the following:

“It is necessary to place the concrete from one side of a blockout so that it flows under and completely fills the area under the blockout. If concrete is placed from both sides, a void or honeycomb is likely to occur beneath the blockout.”

Moving concrete laterally is also needed at the boundary element in shear walls (see photo). The boundary elements are so congested with reinforcing steel that it’s not possible to insert a pump hose or a vibrator from the top of the element. Instead, the vibrator operator is part of the placing crew. The crew member directing the discharge from the pump hose in the shear wall portion adjacent to the boundary element works with the vibrator operator, who vibrates the concrete so it flows laterally into the element. The vibrator operator observes the lateral flow to ensure that the concrete completely fills the element.

If segregation occurs when concrete is moved laterally, it’s most likely to be a problem with the mixture proportions. Segregation causes are discussed in detail in ACI 212.3R-10, “Report on Chemical Admixtures for Concrete.” That document notes, in referring to flowing concrete, that segregation and bleeding can be reduced by increasing the fine-to-coarse aggregate ratio, improving the aggregate grading, or adding other fine material. Entrained air and viscosity modifying admixtures also help to minimize segregation.

What about empirical evidence?

Some ACI documents use empirically based design formulas. The deflection formulas in Chapter 9 of ACI 318 are an example of this approach. Is empirical evidence considered to be as flawed as anecdotal evidence? Not at all. Anecdotal evidence is often used to deduce a conclusion that doesn’t follow from it, usually by generalizing from an insufficient amount of evidence. Empirical evidence is measured, unbiased and replicable.5 The formulas in ACI 318 meet these requirements. They may not predict deflections with great accuracy, but they are based on a lot of measured, experimental data. In short, they have the numbers to back them up.


1. Suprenant, B.A. and Malisch, W.R., Tolerances for Cast-in-Place Concrete Buildings, American Society of Concrete Contractors, St. Louis, MO, 2009, pp. 61-62.

2. Novella, Steven, “The Role of Anecdotes in Science-Based Medicine,” Science-Based Medicine, Jan 30, 2008.

3. Suprenant, B.A. and Malisch, W.R., “How Clean Must Rebar Be?” Concrete Construction, June 1998, pp.517-523.

4. ACI Committee E-703 Concrete Craftsman Series 2, Cast-in-Place Walls, American Concrete Institute, Farmington Hills, MI, 2000, p. 40.

5. Parr, Ben, “Ben Parr’s Entrepreneurial Musings »

Note: ACI documents cited can be purchased at