Why Can Asphalt Laboratory Test Results Be Misleading?

An asphalt mixture may have VMA, VIM, Stability, or other parameters within specification, yet the result can still be misleading if the sample submitted to the laboratory does not represent the actual material or pavement being evaluated.

This is why sampling is a critical part of quality control. Road Knowledge Hub ID treats representative sampling as an important part of understanding laboratory results correctly.

SNI 6890:2014 specifically covers procedures for sampling asphalt mixtures from production units, storage, transportation, and in-place pavement.

The basic principle is:

Non-representative sample → non-representative laboratory result → potentially incorrect quality decision.

What Is a Representative Sample?

A representative sample is a sample that can reasonably reflect the characteristics of the material or work being evaluated.

Imagine there are 1,000 tons of asphalt mixture.

Only a small portion can be sent to the laboratory.

The question is:

Does that small sample actually represent the 1,000 tons?

That depends on how the sample was selected and collected.

If sampling follows the appropriate procedure, the result can provide useful information about the material it represents.

But if the sample is deliberately taken from the best-looking location, the laboratory result may look excellent even though the overall work does not have the same quality.

Why Is Sampling So Important?

Laboratory testing does not begin when the sample enters the testing machine.

It begins much earlier:

Selecting the sampling location
↓
Collecting the sample
↓
Preparing or combining the sample
↓
Identifying the sample
↓
Transporting it
↓
Testing it

An error at the first stage can remain embedded throughout the entire process.

This means that even a perfectly performed laboratory test can produce a misleading conclusion if the sample itself was not representative.

What Is Random Sampling?

Random sampling means selecting units or locations using a process that gives the relevant portions of the population an appropriate opportunity to be selected.

The purpose is not:

“Find the worst location.”

And it is not:

“Find the best location.”

The objective is:

“Select the sample without deliberately choosing material based on the result you want to obtain.”

This helps reduce sampling bias.

What Happens If a Sample Is Deliberately Taken from a Good Location?

Imagine a pavement section with varying conditions:

Good

Average

Poor

But the sample is deliberately taken from the best-looking area.

The laboratory then reports:

VIM → Pass

VMA → Pass

Stability → Pass

On paper, everything looks fine.

But the real question is:

Does the result represent the entire work?

Not necessarily.

This is the danger of sampling bias.

Laboratory Data Can Be Correct, but the Conclusion Can Still Be Wrong

This distinction is extremely important.

Imagine that:

• The laboratory follows the correct procedure.
• The equipment is calibrated.
• The technician performs the test correctly.
• The calculations are correct.

The result shows:

VIM = within specification

There may be nothing wrong with the laboratory test itself.

But if the original sample was not representative, the result is only valid for:

The sample that was tested.

It does not automatically mean:

The entire pavement meets the specification.

Sometimes the problem is not inside the laboratory.

The problem happened before the sample ever reached the laboratory.

A Simple Example

Imagine you want to determine the health of a fish pond.

There are 1,000 fish.

You only select the fish that look healthiest.

The examination shows:

100% healthy.

Does that mean every fish in the pond is healthy?

Not necessarily.

The sampling method was biased.

The same principle applies to asphalt mixtures.

A good sample does not automatically mean the entire work is good.

How Does SNI 6890:2014 Address Sampling?

SNI 6890:2014 covers sampling of asphalt mixtures from several sources, including:

• Production units.
• Storage.
• Transportation.
• In-place material.

For compacted pavement layers, the procedure includes selecting units randomly and taking multiple portions of approximately equal quantity.

Sampling also needs to cover the full thickness of the layer being evaluated without unnecessarily including underlying pavement material.

Therefore, sampling is not simply:

“Take some asphalt and send it to the lab.”

There is a defined procedure behind it.

What About Segregation During Sampling?

Segregation occurs when components within an asphalt mixture become unevenly distributed.

This matters because improper sampling can itself alter the composition of the sample compared with the actual mixture.

For example, if coarse aggregate becomes concentrated in the collected sample, the laboratory may test a material that does not accurately represent the mixture being evaluated.

Therefore, sampling procedures need to obtain a representative sample while minimizing disturbance and changes to the material.

How Many Samples Should Be Taken?

There is no single number that applies to every situation.

Sampling requirements depend on factors such as:

• The amount of material being represented.
• Material variability.
• Project criticality.
• Required confidence level.
• Applicable specifications.

The sampling plan should therefore be appropriate for the population being evaluated and the level of confidence required.

More material or greater variability may require a more carefully designed sampling approach.

Why Is Sample Identification Important?

Even a properly collected sample becomes problematic if its origin cannot be established.

Sample identification and documentation should allow the result to be traced back to its source.

Relevant information can include:

• Project number.
• Road section.
• Material source.
• Production unit.
• Asphalt and aggregate type.
• Sampling location.
• Quantity represented.
• Name and position of the sampler.
• Sampling date.

This is part of traceability.

With proper traceability, the laboratory result can be connected back to:

Sample → Location → Material → Time → Process

Does a Laboratory Result Within Specification Mean the Road Will Be Good?

No.

Laboratory testing is only one part of the quality-control system.

Pavement performance also depends on:

Materials

Mix Design

Sampling

Laboratory Testing

Production

Transportation

Paving

Compaction

Pavement Structure

Drainage

A problem in any of these stages can affect the final pavement.

Can Laboratory Data Be “Arranged”?

Technically, what must be prevented is manipulation or bias in sampling, testing, or reporting.

Examples include:

• Deliberately selecting favorable sampling locations.
• Replacing samples.
• Altering test data.
• Ignoring non-conforming results.
• Using samples with unclear origins.

These practices undermine the integrity of the quality-control system.

But there is an important distinction:

A non-conforming laboratory result does not mean the laboratory failed.

An unexpected result can actually be valuable information indicating that the material or construction process needs further investigation.

💡 Field Note

When reviewing a laboratory report, don’t only look at: VIM ✓ VMA ✓ Stability ✓ Also ask: Where was the sample taken? Who collected it? When was it collected? How much material does it represent? Was the sampling location selected appropriately? How was the sample handled? Is the sample fully traceable? Because even a technically correct test result can lead to the wrong decision if the sample is not representative.

Field Note

🎯 Key Takeaway

Quality control does not begin in the laboratory. It begins with sampling.

A representative sample is essential for ensuring that laboratory results actually describe the material or pavement being evaluated.

SNI 6890:2014 specifically addresses sampling procedures for asphalt mixtures from production, storage, transportation, and in-place conditions.

So don’t only ask:

"Do the VMA and VIM values meet the specification?"

Also ask:

"Does the sample that produced those numbers actually represent the work?"

Because:

Correct data + correct sample = a more reliable basis for decisions.

While:

Correct data + biased sample = potentially misleading decision.

---

Frequently Asked Questions (FAQ)

Glossary