Ah, to find that mean with step deviation, you assume a middle point as your 'a', then reduce all class marks by it, divide by the class width to get your 'u'—smaller numbers, you see? Then it's just Σfu divided by Σf, all multiplied by the class width, and add back your 'a'. In my market research, I use it for monthly grain purchase data, grouping households. But a proper assumed mean isn't always the mathematical middle; I often pick the class with the most familiar, steady spenders. Life has taught me that the truest average grows from the most common ground, not just the centre of the table.
#javelin#throws
Jump in to reply — no account needed.