1.3 Psychophysical Methods

Scientists who study sensation and perception often use psychophysical techniques in their research studies. Psychophysical methods investigate a person’s (psychological) response to a physical stimulus. In other words, how does a physical change in a stimulus affect a person’s perception of it? Psychophysics helps to build a bridge between the physical world and our experience of it. For example, psychophysical methods can help us to determine how loud a sound needs to be before we can hear it or how much sugar we need to add to a cup of coffee to make it noticeably sweeter. Psychophysical methods typically measure people’s sensitivity to sensory stimuli in terms of thresholds. For example, you might describe yourself as someone who is highly sensitive to pain – in other words you have a low threshold for pain. Researchers can measure your threshold for pain with considerable accuracy using psychophysics.

We can measure different types of sensory thresholds, but let’s begin with a common concept in sensation and perception – an absolute threshold. An absolute threshold is the minimum amount of stimulus energy that must be present for a person to detect a stimulus. Another way to think about this is that we can ask how dim can a light be or how soft can a sound be for us to be aware of it The sensitivity of our sensory receptors can be quite amazing. It has been estimated that on a clear night, the most sensitive sensory cells in the back of the eye can detect a candle flame 1.6 miles away (Krisciunas & Carona, 2015) .

Let’s take a closer look at these three different methods, with some specific examples to help you to understand what it would be like to be a participant in one of these studies.

Method of Limits
Let’s start with the Method of Limits. Let’s imagine the experimenter wants to measure the absolute threshold for a small circle of light. In other words, how bright does the light need to be in order for us to just be able to see it? As we can see in the left-hand column of Table 1.1, the experimenter first presents a light with very low brightness (1 unit) and asks the participant whether they can see it . They cannot and so they record a “no response” in the table and they then very gradually increase the brightness, one step at a time until the participant says they can see the light (in Table 1.1 below, this occurs when the light reaches 29 units of brightness). The experimenter then stops increasing the brightness and records the response as a “Yes”. This set of steps (from 1 to 29) is called an ascending staircase because the brightness was increasing a little at a time. The threshold on this staircase is estimated to be halfway between the last “no” response (22 units) and the first “yes” response (29 units) i.e.,  25.5 units of brightness. This is also called the cross-over point. The ascending trial is then followed by a descending trial. In the descending staircase, the light starts off bright (50 units) and is gradually decreased in intensity until the participant says they can no longer see it (15 units). The cross-over point on this staircase is between the last “yes” response – 22 units and the “no” response (15 units).

It is quite common to have higher thresholds on ascending compared to descending staircases. This is thought to be due to errors of habituation. In other words, we get into a “routine” or a habit of giving the same response. In ascending staircases, the habitual response is to say no, this means we tend to go past the actual threshold before we change to a yes response. The opposite is true for descending trials. One way to try and compensate for these errors is to use multiple and equal numbers of ascending and descending staircases and then calculate the average of the thresholds across all the staircases. In the example below we have two ascending staircases and two descending ones – so we can take the mean of four cross-over points. Experimenters also use different starting points on each staircase type to avoid the participant trying to guess when they should change their response – based on the previous staircases. This helps us to avoid these errors of anticipation. In this example,  the first ascending staircase starts on 1 unit of brightness, but the second one (2A) on 8 units.

Method of Constant Stimuli

The experimenter presents very many stimuli of different intensities in a random order and asks the observer whether they can see the light for each presentation. Typically more stimuli are presented than in the Method of Limits. The experimenter records the results for each stimulus intensity as a percentage of the times that the participant says that they saw the light. The experimenter will graph the data as in Figure 1.3. Each circle represents multiple trials at that specific intensity.  For example, as seen in Figure 1.3, at very high intensities (36 units) a participant will say that they see a light all the time (100%) and at very low intensities (e.g., 1 unit) they will say that they see it none of the time (0%). Therefore, the graph is S-shaped with flat ends, the experimenter would use the graph to find the intensity at which the observer says that they see the light half of the time (50%), which is 17.5 in the example in Figure 1.3. Since there is no predictability to how the stimuli are presented, this method eliminates errors of habituation and anticipation. However, this method is slow and you would need a computer (or good graphing skills) to calculate the threshold (50% mark) for each participant.

Magnitude Estimation

There are times when we want to measure more than a person’s threshold -for example, we might want to know a person’s perception of the intensity of a stimulus- we called this magnitude estimation of perceived intensity. Experimenters can measure perceived intensity at different intensity values to see how accurate we are. For example, we are very accurate in estimating the length of a line once we are given a standard measure to compare it to. On the other hand we tend to overestimate the intensity of an electric shock (response expansion) and underestimate the brightness of a light (response compression). A psychophysical researcher named Stanley Smith Stevens asked people to estimate the magnitude of their sensations for many different kinds of stimuli at different intensities, and then tried to fit lines through the data to predict people’s sensory experiences (Stevens, 1967). What he discovered was that most senses could be described by a power law of the form P ∝Sⁿ
where P is the perceived magnitude, ∝ means “is proportional to”, S is the physical stimulus magnitude, and n is a positive number. If n is greater than 1, then the slope (rate of change of perception) is getting larger as the stimulus gets larger, and sensitivity increases as stimulus intensity increases. A function like this is described as being expansive or supra-linear. If n is less than 1, then the slope decreases as the stimulus gets larger (the function “rolls over”). See Figure 1.4. These sensations are described as being compressive. Stevens’ Power Law is useful for describing a wider range of senses.

Both Stevens’ Power Law and Weber’s Law are only approximately true. They are useful for describing, in broad strokes, how our perception of a stimulus depends on its intensity or size. They are rarely accurate for describing perception of stimuli that are near the absolute detection threshold. Still, they are useful for describing how people are going to react to normal everyday stimuli.