I am a chemist and fancy myself as a scientist, a respector of numbers, data and the treatment thereof.

      If I were to undertake a research on this subject, the first thing I would do was get together a group of interested parties and make a list of every conceivable variable. I'm sure one could come up with at least twentyfive possibilities. Then one could collect data that would cover these possibilities. Finally I would put the data through a multi-variate analysis to assign probable contributions of all the various possibilities. Probably 75% of the variables could be eliminated on the basis of too small a contribution to be significant.
      Perhaps a more sophisticated data gathering could then be undertaken based on these results. Of course the data already taken could not be used in this subsequent study.

      It doesn't sound to me that anything approaching this was done. I can think of one variable that must be at least equal in importance if not far surpassing daylight running lights, the color of the car.

      Do I need to point out that most of the studies were done in higher latitudes, Canada, Sweden and such? This of course points out that daylight running lights are important in the twilight and the dawn. In these areas these periods can be a number of hours in length. But in California they are probably about 30 minutes and in Mexico City a few minutes!

      It appears to me that these studies were undertaken with the desired result already determined. This is the absolute worst scenario for research. It doesn't even deserve the name.

      One of the articles seemed doubtful that the data supported the conclusion but ended up saying that why not do it anyway because it doesn't cost very much.

      Another article has the statement, "The present analysis of the Swedish study regarding the Nationwide DRL implementation shows that there is no unequivocal evidence for an effect of DRL on accident rates."

      General Motors of course saw this as an advertising gimmick and jumped on it immediately. I'm sure its effect was visible in the profit and loss statement!
      I have been told the main reason General Motors adopted daytime headlights is that they are required in Canada where they have many hours of twilight and dawn. And General Motors didn't want to have to make two models of everything

      I question whether it is ethical for an insurance company to give a discount to anyone with daylight running lights