Ethnographic data is often regarded solely as qualitative data. And while there is certainly nothing wrong with qualitative data, business folks are prone do dismiss it, largely because of their own epistemological constructs. But that is another topic about which to write. No, this is a more hands-on blog entry about pile sorting. It is about another research tool to add to the ethnographer’s tool kit. Very often quantitative techniques can serve to illuminate patterns and themes in the ethnographic data which would be ordinarily missed. Pile sorting is one technique we should all become familiar with using. A brief description is given here of how to do a Successive Pile Sort using the same stimuli used in the fieldwork is provided.
Pile sorting begins by defining topics, be they images, words, or concepts that are written on index cards or the like. Each of these items or values is derived from data uncovered during participant observation and are relevant in some way to the group of informants.
On the back of the card a letter of the alphabet for identification is also written – this makes coding and mapping easier down the road. Since there are thirty plus items in a complete set, lower case letters of the alphabet obviously have to be used after the twenty-sixth item. Another set of blank index cards is then numbered one to x (the total number of cards in the original set). A complete set of stimuli cards for a Successive Pile Sort procedure consisted, therefore, of stimuli cards with a uniquely corresponding alphabet letter and another set of numbered cards.
The next task is to simply have the informants sort the stimuli cards into various piles of shared meaning. It should be noted that there are variations in the procedure at this point. For example, I typically choose to administer the unconstrained top-down sort version of the Successive Pile Sort because doing so means the informants are not restricted in any way as to how they order the various items. In some versions of the Successive Pile Son, the informant is only allowed to make two piles . In the unconstrained version they were allowed to make as many piles as they wish and include as many items as they wish in each pile using any criteria they see fit.
Typically, most informants will make four to five piles on their initial sorting of the stimuli. Once the participants are satisfied with their piles of items, they are asked to explain why they choose to group certain items together. They are then asked if they want to change anything, such as move a particular item from one pile to another, create another pile or combine two or more piles etc. Typically some degree of shuffling will occur. When this is done, informants are asked their reasoning – it is in the recombination that new meanings and linkages are uncovered. At this point an index card representing (N-1) is placed between the piles.
The card is placed after the last item of one pile and before the first item in the adjacent pile so there is a number between two item cards. Once the piles are separated, the informant is asked to go back to all the original piles and split one of the piles in order to make (N+1) piles. This process continues until each item is separated from another by a numbered index card, or each item forms a pile onto itself. You then record the sequence of items (as identified by alphabet letters) and the numbers. Once the data are collected, they are entered into a computer program, such as ATLASti or ANTHROPAC, and the results are analyzed through multidimensional scaling.
The rationale behind employing this technique is almost as important as the technique itself. The Successive Pile Sort can not be appropriately used to analyze every kind of ethnographic data. By using this technique and expressing it through multidimensional scaling, there is an assumption that there is an internal structure to the items or stimuli being sorted. In the minds of the informants, the items being sorted are related to one another in some way. The Successive Pile Sort serves to reveal the relationship among these items in a way that may not have been previously obvious or conscious to the informants or ethnographer. It can also be used to “map” the linkages between concepts, items, etc.
Ultimately, pile sorting isn’t particularly sexy, but it does add weight to our findings and helps structure complex systems of interaction and meaning. It is another tool to bolster what you do. And another tool to add to the kit is a marvelous thing indeed.