Black Box IEEE 1284 Betriebsanweisung PDF herunterladen (Seite 5)

The Black Box Toolkit Page 5 of 98

1. INTRODUCTION

1.1. Background

Computers, whilst ramping-up in terms of clock speed, are actually no more accurate than

those of a decade ago. In fact quite the reverse can be true with today’s modern multi-

tasking operating systems. Even if you make use of a recognised experiment generator,

there is little assurance that your stimulus and response timings are “millisecond accurate”.

Many packages promise to achieve “millisecond precision”. Unfortunately there is a subtle,

yet important, difference between “accuracy” and “precision”. Millisecond precision simply

means that timings are reported in units of a millisecond – there is no assurance that the

actual timings are accurate!

By using the Black Box Toolkit, or BBTK for short, you can check the presentation and

response timing accuracy of the majority of paradigms in use today. If you are measuring

presentation or response events in units of a millisecond, you should be using the toolkit

as a matter of course.

Achieving the best possible stimulus display timing is becoming more important as

researchers push the envelope with the types of studies they run and data they collect.

Synchrony between visual and auditory materials for example is often prone to larger

variation than many researchers acknowledge. Response timing can also be affected

adversely. The mere act of swapping one response device for another can statistically

alter your results. This is a proven fact – what’s more, without checking you would never

know!

Within any study that has not already been “calibrated” using the Black Box Toolkit there is

almost guaranteed to be one or more sources of uncontrolled timing error; be this within

presentation or response timing. Such error can adversely effect statistical power,

introduce conditional bias, make replication difficult, and lead to spurious effects. This is

before one verifies the paradigm to ensure that it is actually doing what it has been

designed to do. Honest mistakes in scripting can lead to presentation errors that are hard

to detect due to high presentation rates.

By using the BBTK you can help ensure that:

• Your experiment is performing as intended in terms of presentation and

synchrony. For fast presentation schedules it can be difficult for the researcher

themselves spot errors unaided

• You can tune presentation schedules to achieve the best possible presentation

accuracy and consistency (if you don’t know what’s broken you can’t fix it!)

• You know what the absolute error and variance is within your chosen response

device – remember these can vary enormously! Armed with this knowledge you

may decide to change device or perform a post-hoc statistical correction

• You improve your chances of replication and internal consistency

• Above all you improve the quality and respectability of your research

1 2 3 4 5 6 7 8 9 10 ... 97 98

Keine Kommentare

Black Box IEEE 1284 Betriebsanweisung Seite 5