Dear all,
I am processing nondiagnostic pottery shards. I analysed fragmentation and abrasion in the field.
I first divided my nondiagnostics into clay groups. After that for each clay group I sorted out the individual shards according to size categories using a grid (1x1cm, 2x2cm, 3x3cm etc) and I also analysed shards from each clay group according to 3 levels of abrasion.
Now, I want to calculate the level of fragmentation and abrasion of each clay group. Any suggestion on what would be the best way to do this?
Best wishes,
Uros
It is possible to do Custom tables analysis with Chi square to separate
The parameters in line can be clay, size. Then put the response in column
This can easily be done using SPSS
A Chi-squared test, Gangue Tiburce , will not be aware of the ordered nature of the categories. For this reason, the test will fail to detect a consistent increase in abrasion across the size categories.
A correlation such as Kendall's tau-b or an ordered logistic regression would be more appropriate. The advantage of the logistic regression model is that you can look at multiple predictors.
Dear Uros,
if you are looking at Fragmentation of pottery per different clay Groups, I would suggest to record weight instead of size. It it easier to measure and you don't have to define Groups (1x1 cm, 2x2 cm) in adavance. I include a paper on Analysis of sherds by weight on medieval pottery from field Walking, which will be published at the end of the year (unfortunately it is in German)
Best wishes
Christoph
Dear Christoph,
Thank you! I wanted to investigate if the degree of fragmentation and abrasion is depending on the context, that is why I measured the size of the non diagnostic sherds.
Best wishes,
Uros
- Badges
- Science topic