How to carry out pile sorting and how to analyse the data with Anthropac: a tutorial
What is pile sorting?
Pile sorting is a method used to understand people’s perception and structure of a cultural domain through an observation of how people classify and group the items of the domain studied (Bernard 2002). For example, if a researcher wants to know how Germans perceive different types of food plants, he could investigate how they classify them in different groups like vegetables, fruits and spices.
How to carry out pile sorting
The researcher gives a set of cards/items to each informant and then asks the informant to divide the cards/item by groups. In order words, informants are asked to pile similar items together. For example, when investigating the classification of plant food the researcher could use either a set of cards with the name of the plants printed on a side, either photos of the vegetable, either the actual vegetables and fruit.
The researcher explains that groups can have as many cards as an informant wish. When the informant finishes sorting the cards/items, the researcher asks to the informant to make subgroups if this is possible. The researcher also asks the respondent to explain the criteria used to sort the cards/items and to name each group formed.
If the exercise is done with cards, the researcher can either write the names or print the pictures of the items to be sorted. Obviously, if the informants are illiterate writing the names is not a viable option and photos or physical items must be used.
How to analyse data collected through pile-sorting
The data collected through pile sorting is usually entered into a text file according to a specific format (see image below or Download example ) and then analysed with the software ANTHROPAC (Borgatti 1996).
To carry out pile sorting analysis on ANTHROPAC select DATA>IMPORT>PILESORT and then type the name of the text file with the pile sort data and press F10
- The outputs of this analysis are:
- A matrix which describe, at the group level, how similar are each pair of items (item-by-item matrix) that have been sorted (AGPROX).
- Item-by-item matrix for each informant describing how each informant grouped the items (INDPROX).
- ANTHROPAC automatically computes the correlation between the individual matrix and the aggregate matrix and store the results in file CORR.
- Look at the file with the results
- Open the folder where you are working for example C:\folder
- There should be a file named OUTPUT.LOG open it with wordpad, it shows the results of your pile sorting analysis
- Further analysis of the data:
- AGPROX can be used as input data to obtain a graphical representation of the similarities among items. This graphical representation can be useful to identify clusters of item perceived as similar.
- INDPROX can be used as input for consensus analysis to assess how much consensus there is among informants.
FIGURE1: example of file with pile sorting data, Download
Bernard HR. 2002. Research methods in anthropology. Altamira Press, Walnut Creek, CA.
Borgatti S. 1996. ANTHROPAC 4.0. Analytic Technologies, Natick, MA.