You know what really bothers me when I read Ofsted reports. Think about just how critical a tiny number of lesson observations are to an inspection, to the extent that it’s considered legitimate to refer to parts of single lessons in justifying judgements on teaching and learning. Put this in the context of what Robert Coe said at ResearchEd 2013 about the number of lesson observations required to approach a reliable judgement and I think, even if Ofsted were to significantly raise their game on the quality assurance of observations (and observers), the quality of teaching and learning judgement is always going to be influenced by a few teachers having particularly good or bad days, or even a few inspectors having ‘outlier’ days, and that just can’t be right. I imagine this is obvious to everyone, except that if it was obvious to Ofsted, they surely would see that quoting single instances of perceived good or poor practice emphasises the tendency to extrapolate a small amount of observation across an entire school. If I were writing a report I would be trying to give the impression that the judgement emphatically was not swayed by a few particularly great or grotty lessons, not trumpeting them as some kind of deal clencher.