Survey Says!

The blind cable listening test survey closed Thursday, and  I announced some preliminary results yesterday in my session at the USITT conference.

Here is what you heard in the audio file:

X01: Conventional Cable

X02: Conventional Cable

X03: Star Quad Cable

X04: Conventional Cable

X05: Star Quad Cable

X06: Star Quad Cable

X07: Star Quad Cable

X08: Conventional Cable

X09: Star Quad Cable

X10: Star Quad Cable

Clip Y was Star Quad; Clip Z was conventional.

In ABX testing, to achieve a statistically significant result, it's commonly thought that listeners need 8 or more correct guesses as to what "X" was.  

I had 78 survey respondents, and 3 got 8 or more correct guesses.  So, given the criteria above, only 3.8 % of my survey respondents were possibly able to tell the difference between the two kinds of cables in our experiment.  I say "possibly" because it's conceivable that those three people got the right answers due to chance (in fact, one of the three is one of my students, who has significant hearing damage).

¡

Will be doing full statistical analysis later.

¡

For now, I’m reporting

the number of people who got 8 or more correct guesses (a 95% confidence interval—a measure of confidence in the results).

Intuitively: there are many valid random guess possibilities to give 3-7 correct answers,

but very few that give 0-2 or 8-10 correct