The psychological interaction of spam email features

Abstract

This study explored distinct perceptual and decisional contributions to spam email mental construal. Participants classified spam emails according to pairings of three stimulus features - presence or absence of awkward prose, abnormal message structure, and implausible premise. We examined dimensional interactions within general recognition theory (GRT; a multidimensional extension of signal detection theory). Classification accuracy was highest for categories containing either two non-normal dimension levels (e.g. awkward prose and implausible premise) or two normal dimension levels (e.g. normal prose and plausible premise). Modelling indicated both perceptual and decisional contributions to classification responding. In most cases, perceptual discriminability was higher along one dimension when stimuli contained a non-normal level of the paired dimension (e.g. prose discriminability was higher with abnormal structure). Similarly, decision criteria along one dimension were biased in favour of the non-normal response when stimuli contained a non-normal level of the paired dimension. Potential applications for training are discussed. Practitioner summary: We applied general recognition theory (i.e. multivariate signal detection theory) to spam email classification at low or high levels of three stimulus dimensions: premise plausibility, prose quality, and email structure. Relevant to training, this approach helped identify perceptual and decisional biases that could be leveraged to individualise training.

Keywords: Spam email; phishing judgment; spam attention; spam features; spam judgment.

Figure 1.

Sample email messages displaying at least one non-normal dimension. Top: awkward prose; Middle: abnormal structure; Bottom: implausible premise.

Figure 1.

Sample email messages displaying at least one non-normal dimension. Top: awkward prose; Middle: abnormal structure; Bottom: implausible premise.

Figure 2.

Hypothetical contours of equal likelihood from factorial combination of features along two dimensions. The left panel shows a violation of PI because there is a nonzero covariance within a category. The middle panel shows a violation of PS along the A dimension due to a difference in perceptual effects at levels B1 and B2. The right panel depicts a violation of DS along dimension A since the decision bound is not orthogonal to the coordinate axes.

Figure 3.

Classification accuracy on each category from the dimension pairings used in the experiment. Error bars show standard error of the mean.

Figure 4.

Best-fitting configuration of perceptual distributions from block 1 (Prose x Structure) trials. Ellipses are contours of equal likelihood. Symbols represent category means, and brackets indicate distance between category means within perceptual space. NL = normal, AW = awkward, AB = abnormal. In the legend, the dimension on the x-axis (prose) is listed first in the ordered pairs. The dimension on the y-axis (structure) is listed second in the ordered pairs. Filled shapes represent a higher (non-normal) level on the y-axis, while triangles represent a higher (non-normal) level on the x-axis.

Figure 5.

Best-fitting configuration of perceptual distributions from block 2 (Premise x Prose) trials. Ellipses are contours of equal likelihood. Symbols represent category means, and brackets indicate distance between category means within perceptual space. NL = normal, IM = implausible, AW = awkward. In the legend, the dimension on the x-axis (premise) is listed first in the ordered pairs. The dimension on the y-axis (prose) is listed second in the ordered pairs. Filled shapes represent a higher (non-normal) level on the y-axis, while triangles represent a higher (non-normal) level on the x-axis.

Figure 6.

Best-fitting configuration of perceptual distributions from block 3 (Premise x Structure) trials. Ellipses are contours of equal likelihood. Symbols represent category means, and brackets indicate distance between category means within perceptual space. NL = normal, IM = implausible, AB = abnormal. In the legend, the dimension on the x-axis (premise) is listed first in the ordered pairs. The dimension on the y-axis (structure) is listed second in the ordered pairs. Filled shapes represent a higher (non-normal) level on the y-axis, while triangles represent a higher (non-normal) level on the x-axis.