You write as though you thought there *OUGHT* to be a relationship between weight and crash test ratings. Why should there be? Your data seem to be trying to tell you that the correlation is close to 0 for front ratings. Virtually all cars these days have some kind of fore-and-aft crash protection -- seat belts, air bags, engine compartments designed to crumple gradually rather than deposit the engine in the driver's lap -- and these are devices whose mass (and effectiveness) are not much related to the mass of the vehicle.
Side ratings are another matter. Not all cars offer side air bags, or structural protection against side impacts. One would therefore expect some correlation with weight, since such devices do cost something in terms of weight, and may possibly be more likely to be found in more luxurious (and therefore, on average, heavier) vehicles. You report a correlation of 0.6, which sounds reasonable enough for a correlation essentially between weight of vehicle and presence/absence of side impact protection. You may also find it illuminating to consider exactly what information goes into those ratings, and to ask yourself what it might mean if there WERE a correlation between weight and rating for frony/rear impacts. On 1 Jun 2003, Serge wrote: > I have to do a final project for my Statistics class. I chose to do > a project on the relationship between a car's weight and its 4 crash > test ratings (2 front for passenger and driver and 2 side for front > and rear). The ratings are 1 - 5 (the number of stars); the higher > the rating, the safer the car. I got all my data from crashtest.com. > In all, I got data for 187 cars - that is, the weight and the 4 > ratings. However, when I did a scatterplot of any 1 of the front > ratings vs. weight the r coefficient was very close to 0 and the > plot did not show a pattern. When I plotted the side ratings vs. > weight, r was about .6 in both cases (a mild relationship). > Thus, I can't really conclude anything from this. Why not? You appear to be able to conclude that there is no correlation visible between weight and front-impact ratings, and that there is a positive correlation (have you looked to see whether 0.6 differs significantly from zero? I suspect you'll find it does) between weight and side-impact ratings. > What I am thinking of doing and what I am asking about is > this: If I separate my data into 9 groups based on weight where each > group spans 500 pounds so that group1 has cars with weights 1501 - > 2000 and group > 9 has 5501 - 6000 and then take the mean of all 4 ratings in each of > these groups and plot them against the mean weight in each group for > all 4 ratings (again, 4 separate plots), will I get a more definite > relationship (I haven't done it yet so I don't know) and even more > importantly, will this apparent relationship be statistically > significant and why or why not (why does grouping help or why not)? If > not, can you suggest something else that I may be able to do. Why should such an effect, if it could be conjured up, be a "help"? Are you operating under the misperception that research is only valuable if it leads to "significant" results? That surely is not what your statistics instructor is trying to teach you (if it is, the person is incompetent and should be fired). The point of doing research is to try to find out what IS, not to justify one's personal prejudices. As Einstein is alleged to have put it (I've seen this on a .sig of one of our colleagues, don't remember who), "If we knew what we were doing, it wouldn't be research." ----------------------------------------------------------------------- Donald F. Burrill [EMAIL PROTECTED] 56 Sebbins Pond Drive, Bedford, NH 03110 (603) 626-0816 . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
