Code as ‘Other’: Reducing the Human Experience to a Number

“He came into my house.” == ‘Other’
“He kidnapped me.” == ‘Other’
“I was raped.” == ‘Other’

I recently had the opportunity to run a study on polygamous relationships in rural regions of Eastern and Western Uganda. For each interview with the male respondent and each of his wives, the survey sought to better understand the dynamics of each respective relationship as well as the reason behind each marriage, including how each couple met.

Deep in the daunting task of coding responses that the enumerator recorded as ‘Other’, the dark side of quantitative data collection reared its ugly head.

This was particularly the case for the question “how did you meet?” Many responses that were noted as ‘Other’ actually fit within a designated category for which we had a coded number (ex: “we lived in a village together” receives the code 3, which signifies there was no formal introduction between husband and wife; “he was my friend’s brother” becomes the code 1, meaning introduction via friends). A simple read-replace file makes the switch.

The problem came when I encountered the above responses. “He came into my house…” “He kidnapped me.” “I was raped.” According to the process at hand, each would be a 3 – no formal introduction.

When we code these responses, then the experience of a woman who has been kidnapped, or raped, is relegated to simply a normal meeting between a man and a woman that resulted in marriage. To those who will analyze this data, had I made the coding switch, these respondents would be just like thousands of others, bereft of personal struggle, of personal trauma, of personal tragedy.

I will admit publicly to any PI that looks over this data – I could not make the switch. These stories deserve to be heard, particularly if the woman had the strength to tell them to a total stranger when our enumerators made their way to her village.

This is perhaps my biggest qualm with quantitative data. While it is imperative to gather statistics to draw inferences about populations in order to design interventions that serve them best, this very act leaves out countless stories that demand as much attention (if not more) as the proposed intervention at hand (in this case, an incentive program for men to include their wives on their land titles).

How many women have used these types of interviews as an opportunity to tell a deeply powerful and traumatic story, only for it to be lost in a sea of numbers because their response did not fit within the indicators the question was seeking to measure?

How many stories are falling through the cracks? How many cries for help are going unheard?

I’m not saying quantitative data is harmful (necessarily), although it pains me to know that these women’s stories will do nothing to bring these men to justice. Perhaps I’ll switch over to qualitative soon…