While the Edinburgh Handedness Inventory has been widely used, there have been few studies assessing its factorial validity. There is evidence that the original instructions and response options are difficult to understand. Using simplified instructions and response options, the Edinburgh Handedness Inventory was administered on a sample of 1514 participants using an online questionnaire. In accordance with previous research, a model of the 10-item inventory had poor fit for the data. This study also detected model misspecification in the previously-proposed 7-item modification. A 4-item Edinburgh Handedness Inventory - Short Form had good model fit with items modelled as both continuous and ordinal. Despite its brevity, it showed very good reliability, factor score determinacy, and correlation with scores on the 10-item inventory. By eliminating items that were modelled with considerable measurement error, the short form alleviates the concern of the 10-item inventory over-categorising mixed handers. Evidence was found for factorial invariance across level of education, age groups, and regions (USA and Australia/New Zealand). There generally appeared to be invariance across genders for the 4-item inventory. The proposed Edinburgh Handedness Inventory - Short Form measures a single handedness factor with an inventory that has brief and simple instructions and a small number of items.