Data work in schools tends to rely heavily on popular software tools such as Microsoft’s stalwart ‘Excel’ spreadsheet. One might like to imagine that software-driven analysis leads teachers to conform to statistical rules and procedures. Yet, rather than acting as a standardized corrective to ‘sloppy’ use of data, our research is highlighting how statistical software packages and spreadsheets such as Excel allow all sorts of spurious and statistically inappropriate ‘calculations’ to be run regardless of their robustness.
Of course, schools are not alone in this. The consequences of over-reliance on Excel can be extreme. Take the recent story from the UK that Public Health England inadvertently left off nearly 16,000 COVID test results from the official daily figures because the .XLS files being used had exceeded their file-size limit. This error was estimated to result in UK contract tracers ignoring up to 50,000 potential COVID carriers.
In their reporting of this mispractice, the Guardian offered a number of other examples of the ‘incredibly fragile’ (Kwak 2013) nature of Excel data analyses. This includes the investment bank JPMorgan incurring a $6bn loss after an Excel spreadsheet cell was erroneously set up to divide by the sum of two interest rates, rather than the average. Elsewhere, the Human Gene Nomenclature Committee has been forced to change the names of 27 genes due to Excel’s continual autocorrecting and mis-formatting of original names such as SEPT1 and WARS.
All told, it is important to consider the ways in which software packages such as Excel are significant actors in the circulation of school data – shaping the ways in which data is produced, analysed and used.