Time-saving bias is the tendency to misjudge how much time a change in speed will save or cost. In research tasks, people commonly underestimate the gain from increasing a low speed and overestimate the gain from increasing an already high speed.

The trap is mathematical. Speed rises in a straight line, but the time needed to cover a fixed distance does not fall in a straight line. A larger-looking speed increase can therefore buy less time than a smaller increase made from a slower starting point.

A fictional shuttle calculation

This example is wholly invented. The museum, shuttle route, team, options, and decision are not study data. The arithmetic assumes a fixed 12-kilometer distance and a constant mean speed.

A museum operations team is comparing two idealized shuttle segments. On the slower segment, raising the mean speed from 24 to 30 kilometers per hour cuts the calculated trip from 30 minutes to 24 minutes. That is a six-minute saving.

On the faster segment, raising the mean speed from 60 to 66 kilometers per hour cuts the calculated trip from 12 minutes to about 10 minutes 55 seconds. That saves about 65 seconds.

Both changes add 6 kilometers per hour. The increase from the lower starting speed saves more than five times as much time.

That result says nothing about what the fictional team should do. Actual shuttle schedules include stops, dwell time, traffic, road design, speed limits, and safety rules. The example isolates one calculation so the bias is easier to see.

The reciprocal math behind time-saving bias

For a fixed distance, duration is:

time = distance / speed

If the original speed is V1, the higher speed is V2, and the distance is D, then the idealized time saving is:

time saved = D × (1/V1 - 1/V2)

When distance is measured in kilometers and speed in kilometers per hour, this result is in hours. Multiply by 60 to convert it to minutes.

The reciprocal terms matter. At low speeds, a small change removes a relatively large share of the original duration. At high speeds, the curve has flattened, so each additional unit of speed removes less time.

This is why the starting speed cannot be treated as background information. For a fixed 10-kilometer trip, increasing from 30 to 40 kilometers per hour saves 5 minutes. Increasing from 70 to 110 kilometers per hour saves about 3 minutes, even though the second speed increase is four times larger. Ola Svenson used comparisons of this kind to study decisions between time-saving options. (Svenson, 2008)

What the experiments show

The evidence supports a recurring directional error, but it does not supply a universal rate for how often people make it.

In one three-study investigation, Eyal Peer tested several forms of the problem. Study 1 included 79 licensed active drivers. Participants estimated journey duration after a speed increase, distance covered in a set time, or the speed required to meet a shorter duration. Their answers showed the expected errors across those broad formats. One unusual item involving a cheetah over a very short distance produced a reversed pattern, a useful warning that framing and scale can matter. (Peer, 2010)

Study 2 included 139 licensed active drivers. It compared questions about remaining journey time with questions about time saved, and numeric responses with a visual scale. Across three low-speed problems, average underestimation ranged from 6% to 38%, with a median of 15%. Differences between the four presentation conditions were inconsistent. The main directional finding survived, but its measured size depended partly on the task. (Peer, 2010)

A separate hypothetical-choice study asked drivers about a 20-kilometer journey. On average, participants estimated that increasing from 40 to 50 kilometers per hour would save about 4.5 minutes; the correct idealized answer is 6 minutes. Participants classified above the median on the study's bias measure also indicated higher required and preferred speeds and exceeded the stated limit more often in their responses. This is evidence of an association in hypothetical situations. It does not prove that the bias causes actual speeding or crashes. (Peer, 2010)

There is no single settled shortcut

It is tempting to explain every wrong answer with one linear rule. The research is less tidy.

Svenson's early account proposed a proportion-based shortcut: people appear to compare the size of the speed increase with the higher speed. Later studies found that a percentage or difference rule, based on the increase relative to the starting speed, classified more non-normative answers. Participants higher on a need-for-cognition measure also answered correctly more often in one study. (Peer and Gamliel, 2012)

Neither rule explains every error. In the pace experiments described below, many wrong answers fit neither proposed heuristic. A careful summary is that people often fail to use the reciprocal speed-duration relation. Different prompts and different people can produce different wrong paths.

Why the evidence does not make speeding rational

A calculated saving is only one part of a driving decision. The formula does not include collision risk, stopping distance, legal limits, traffic variability, fuel use, or the effect of signals and stops. It also assumes a stable mean speed across the same route.

Real travel often breaks those assumptions. A car that briefly reaches a higher speed may have nearly the same door-to-door time because it encounters the same congestion or red lights. Conversely, a route change can alter distance as well as speed, so a speed-only comparison is incomplete.

The research finding is therefore a reason to check the arithmetic, not permission to accelerate. No idealized time gain makes an unsafe or unlawful speed acceptable.

Can time-saving bias be reduced?

One promising method is to convert speed into pace: the time required per fixed unit of distance. Drivers usually see kilometers or miles per hour. Pace turns the same information around, such as minutes per 10 kilometers.

In a 2013 article, Peer and Eyal Gamliel compared pace and speed formats. In their second study, 68 online participants answered three sets of questions. Participants given pace information were more accurate than those given speed information when estimating the duration of a 10-mile trip (90.81% versus 68.39%), the speed needed for a 10-mile trip (80.00% versus 49.29%), and the speed needed for a 30-mile trip (38.79% versus 20.37%). (Peer and Gamliel, 2013)

The last comparison also shows the limit: fewer than two in five pace-condition responses were correct. Pace helped in those tasks; it did not guarantee accuracy.

Simulator evidence points in the same direction. Gabriella Eriksson and colleagues asked participants to drive the same distance twice and adjust their speed to save exactly three minutes. A control group saved more than the target after starting at 30 kilometers per hour and less than the target after starting at 100. Participants using an inverted-speed display were closer to the target. The study supports a display aid in a controlled task, not a claim that a quick mental reminder permanently removes the bias. (Eriksson et al., 2015)

Sources