A&C_IADI Logo
HOME ABOUT VOLUMES SEARCH

Volumes / Journal / 19 - 2008 / Baxter M., Cool H.

Notes on the statistical analysis of some loomweights from Pompeii

Click on any image to open slideshow

Fig. 1 – Examples of the ceramic loomweights discussed. Diagram A (not to scale) shows how the height was measured using the offset method.

Fig. 1 – Examples of the ceramic loomweights discussed. Diagram A (not to scale)...

Fig. 2 – Two histograms using different bin-widths that show the distribution of weights of loomweights from Pompeii. That to the right suggests the distribution is bimodal.

Fig. 2 – Two histograms using different bin-widths that show the distribution of...

Fig. 3 – An estimated two-component normal mixture for the weights of the loomweights.

Fig. 3 – An estimated two-component normal mixture for the weights of the loomwe...

Fig. 4 – A “pairs” plot for six variables that characterise the loomweights, showing all possible bivariate plots. The upper triangle of plots is the same as the lower triangle, except that axes are interchanged. Treating the base and apex of the weights as rectangular, Topmax is the length of the larger sides of the rectangle at the apex and Topmin relates to the smaller side. Bottommax and Bottommin refer to similar dimensions for the base.

Fig. 4 – A “pairs” plot for six variables that characterise the loomweights, sho...

Fig. 5 – A histogram for the heights of the loom weights with a kernel density estimate superimposed.

Fig. 5 – A histogram for the heights of the loom weights with a kernel density e...

Fig. 6 – A plot of height against weight, with cases labelled by the classification suggested by the mixture analysis.

Fig. 6 – A plot of height against weight, with cases labelled by the classificat...

Fig. 7 – Different ways of displaying the relationship between height and weight. The raw data is to the upper-left; an image plot is to the upper-right; a perspective plot is to the lower left; a contour plot is to the lower-right.

Fig. 7 – Different ways of displaying the relationship between height and weight...

Fig. 8 – A customised version of the contour plot shown in Fig. 7.

Fig. 8 – A customised version of the contour plot shown in Fig. 7.

Fig. 9 – Assuming that the height/weight data can be modelled using a mixture of bivariate normal distributions, this figure shows different ways of displaying the results.

Fig. 9 – Assuming that the height/weight data can be modelled using a mixture of...

Fig. 10 – A plot of the maximum to minimum ratio of the tops of the loom weights against a similar ratio for the bottom of the loomweights.

Fig. 10 – A plot of the maximum to minimum ratio of the tops of the loom weights...

Fig. 11 – A plot of height against weight, labelled according to the shape typology suggested by Fig. 10.

Fig. 11 – A plot of height against weight, labelled according to the shape typol...

Fig. 12 – Similar to Fig. 6 but using all the data and with a modified size classification.

Fig. 12 – Similar to Fig. 6 but using all the data and with a modified size clas...

Fig. 13 – A graphic that suggests that Type 1 loomweights, with square tops and bottoms, are disproportionately likely to be of a heavier weight.

Fig. 13 – A graphic that suggests that Type 1 loomweights, with square tops and ...

Fig. 14 – A plot of weight against base area, for the modified data that omits unusual data. The solid line is a fitted linear regression used to investigate how well base area can predict height if the loomweight is incomplete. The dashed line is a non-linear smoother. It is close to the regression line, suggesting that more complicated methods of prediction are not required.

Fig. 14 – A plot of weight against base area, for the modified data that omits u...

Fig. 15 – A plot of the modified dataset labelled with the Casa del Chirugo groups summarised in Tab. 1.

Fig. 15 – A plot of the modified dataset labelled with the Casa del Chirugo grou...

Baxter M., Cool H. 2008, Notes on the statistical analysis of some loomweights from Pompeii, Archeologia e Calcolatori 19