Casio fx-85GT calculator

9.9.2013 | 16:17



The fx-83/85GT calculator is obviously designed to match school classroom learning programmes.

It is powerful and easy to use, with clear and legible display, and logical entry modes for a range of functions, such as logarithms to any base (though there are also convenience log keys for both base 10 and natural logarithms), powers, and roots. Trigonometric and hyperbolic functions are easily accessible, the latter through a menu on screen. There are nine memory registers, eight of which (A-F,X,Y) are `scratchpad’, for storage only, and the ninth (M) is addressable for addition and subtraction. There is a table function, which can be used to tabulate y-values of an input function for a range of x-values. The table is not addressable, being read-only, and its value is limited, though could be useful in finding, for example, the approximate zeros of polynomial functions.

The fx-83/85GT has a number of useful features for statistical calculations, from random number generation, permutation and combination calculations, 1- and 2- variable statistics, and a range of useful regression types, for fitting curves or straight lines to a series of data points, though no probability distributions. What should have lifted this calculator above the ordinary, or even the very good, unfortunately, has instead let it down with a thud.

I chose this calculator on what seemed to be a promise of useful and practical statistical features listed on the packaging. Especially attractive was the prospect of a list-based statistics data editor. On more conventional calculators, data pairs are entered, and sums and sums of squares accumulated on a rolling basis, for use in calculating parameters like standard deviations, variances, and so on. A list-based editor offers the prospect of being able to easily edit individual data points for recalculation of parameters. This is especially attractive if one is carrying out curve-fitting, say by linear or power or logarithmic regression.

Memory available for data points is limited. For 1-variable statistics, up to 80 values can be entered. For 2-variable statistics (e.g. x,y pairs), this reduced to 40 data value pairs. For school exercise purposes, this might be adequate, but it’s certainly a limitation for more practical applications. Such limitations are manageable. But the worst is yet to come, and is surprising.

Suppose, from an experiment or series of observations, you think there is a linear relationship between your dependent and independent variables (x,y pairs). You carry out a linear regression, having entered your set of data points, to fit an equation of the form y = A + Bx to your data. By inspection, you then figure the relationship is not linear after all, and suspect that a power regression curve might be more appropriate. The moment you leave the fx-85GT statistics editor, and select a different regression model to work with, all your entered data is immediately erased, without warning! Choose a different regression analysis, and you must re-enter all your data all over again. Calculated regression coefficients are not stored in any of the scratchpad memories either, so leave the stat editor, and these too are lost. Every time you leave the stats editor, this happens. Data entered cannot be used for re-modelling under a different regression analysis.

This is an extraordinary weakness in design, and wipes out a major rationale for having a list-based stats editor, which is the ability to re-interpret or re-model the same data in different ways. The stats list is not addressable either, meaning you cannot use an individual data point in a further calculation. All you can do is edit it.

If you need a calculator with good statistical features, the fx-83/85 may not be your best choice, despite its promotion of a list-based stats editor. Although the cost is higher, a far better calculator is one from the Texas Instruments TI-83 or TI-86 family, which have good list-based statistics editors and facilities, and which leave the fx-83/85GT stumbling on the starting blocks. Otherwise, this is an agreeable and well thought out calculator, with much to commend it.

No comments

Leave a Reply

Your email address will not be published. Required fields are marked *