Publication bias and confounders in product reviews - TAS, Stereophile, Audiogon, etcetera


Folks-

Since I am a research professor at a major medical school in the U.S., I am used to identifying and using statistical measures of such bias in scientific research.

In Japan, I have read that a product reviewer who writes for magazines or websites are paid fees by manufacturers. I have noted that a similar thing may be happening here in the U.S., both reading TAS, Stereophile, etcetera, as well as noticing comments from individuals on this and other websites, many of whom are also dealers of these products.

As an example, I am somewhat of a computer nerd and have been downloading high-resolution audio files for almost a decade. That being said, I have been looking to buy a relatively high-end SACD player for my large collection of CDs and SACDs. I have noted the following:

1. There are few-to-no reviews of DCS players (e.g., Puccini SACD player, somewhat outdated but can be upgraded) and almost no published U.S. reviews of the Marantz SA-10 SACD player that was released about a year ago. In contrast, SACD/CD players including those from Esoteric, Hegel (CD only), Ayre, PS Audio, MBL, and other brands commonly appear in formal reviews, which are all favorable. Does this mean that products which have been reviewed but which are not well-liked by reviewers are not published?;

2.  Comments in this and other forums mention that one or another SACD player or other product "must not be that good because they appear often as used equipment for sale..." or something to that effect. This observation may be valid, but could easily be confounded by the number of such products that were, or are, available for sale. The greater the number of products, the greater the likelihood they will appear as used items for sale - it says nothing about the quality of the product. I like to call this the "Ferrari effect", as this manufacturer intentionally limits the number of cars of any model for sale, and the company often only sells to individuals of affluence and/or have purchased cars from them in the past, artificially inflating the value of these cars; 

3. Odd statements about the interesting MQA file format, part of a larger problem of a lack of objectivity in the audiophile community. Recently I read in a publication - "MQA is to conventional audio what quantum mechanics was to classical mechanics" - Really? Does this individual know anything about physics? Or am I taking this all too seriously?

I guess I am asking about the degree of bias in these reviews, to what extent are products reviews influenced by the manufacturers and dealers, and where is the objectivity in this domain?

Thanks for listening to my ranting...Gerry 
gerryah930

Showing 2 responses by gerryah930

My thanks to all of you for the responses, since I am naive in this domain, and expected audiophile magazines might actually provide unbiased product reviews. 

One last tidbit of information, since we routinely perform high-throughput data analysis on billions of sequence reads - the name "Diana Krall" is significantly over-represented in the discussions of musical content between 2015-2017, both on Audiogon and in TA and Stereophile magazines. I guess she is a popular artist and was also once a student at Berklee in Boston (like I was, a long time ago!) - Rip Van Winkle, aka Gerry.


I am not trying to insight a so-called flame war here, but I have to agree with willemj.

If audiophile magazines are "hobbyist" journals, then I would stop the product reviewers from including what appear to be plots of measurements of impedance versus phase, spectral decay, and amplitude versus frequency. I enjoy looking at these graphs but in almost all cases, the reviewers will only hint at issues (e.g., speaker impedance spikes or drops), but the review will end with statements such as "best speakers I have heard, so I bought the review pair" and so forth. I can only believe that there is some economic incentive that is not disclosed.

Comparisons to a hobby and a profession:

1) I used to be involved in Porsche club racing, so I would read porschephile and generic car magazines. This would certainly be considered a wealthy person's hobby, but the car magazines would provide head-to-head comparisons of performance between brands and would make disparaging statements about cars that were poor performers. I have never read head-to-head comparisons in a given category of hardware in any high-end audiophile magazine (e.g., different speaker brands in the same room at the same time, driven by the same audio hardware);

2) In biomedical research, there are many problems and challenges, but to suggest that our collective intent is to do anything but try to arrive at so-called "ground truth" is not accurate. That does not mean there is no publication bias (see [1]), but at least there is recognition that this issue exists, and many attempts to deal with the problem.

- Gerry

[1]  A. César-Razquin et al. A call for systematic research on solute carriers. Cell. 162 (3), (2015), 478-487. https://www.snijderlab.org/pdf/26232220.pdf