does correlation show cause and effect.

Many people me being one of them can get pretty confused when told that the relationship between two factors doesnt show cause and effect.

Surely if more ice creams are sold on hot days then it shows that the heat makes people buy ice creams to keep themselves cool….right?.. unfortunatly not. all it shows is there seems to be a relationship between the two things. but as for causeation it could be because there happens to be a sale on ice creams on the same day it just happens to be hot.

Also if we were to infact use correlation as a way of finding the cause of something, we would be able to do it with any to factors. this would ofcorse mean that wild theories would be made like everyone who gets a girlfriend buys toilet rolls. this would lead to mass amounts of toilet rolls being bought by single guys. when really its just a case of everyone buys toilet rolls not just people with girlfriends.

in conclusion its clear that correlation is too simple as it cuts out to many factors to find causation.


9 responses to “does correlation show cause and effect.

  1. Good blog on an issue that many get confused with. In school I was also told the ‘the hotter the weather, the more ice cream sold’ story as an example of correlation, however, this then lead me to believe that this was causation… as the temperature had caused the increase in sales of ice cream. However causation and correlation are completely different things as you have pointed out, the sales of the ice cream could be due to a completely different factor that has not be pointed out. Correlation simply points out that there is a relationship between the two things; causality goes onto show exactly what the cause of the relationship may be. Showing that the two are not interlinked. A really good site that helped explain everything to me is It uses real life examples to explain everything well and explains that causation sets out to show us the explanation of the relationship whereas correlation simple shows that there is a relationship present.

  2. Pingback: Comments for my TA- due 14/03/12 « raw2392

  3. yes causality is diffficult to detect in most cases and jkind of leaves a chicken/ egg senario. in some studies causality when statistics are used is easy to detect as there is a direct correlation between the two entities which are being tested. In this study there is a very clear causal correlation between homocysteine and particularly heart and pulmonary disease. In some cases the causality between two entities is not as clear thus blurring the finding of a particular field

  4. hey good blog! i think the ice cream story is a classic in describing correlation. perhaps you could have gone into more detail and pushed your point more, there is loads of resources online discussing this topic. I found a good paper that does the job
    Also i think you could have used an experiment to back your blog further and included more information. causality and correlation are two different things! you could have included how they affect the general public (how people try to get customers to buy more) for example raising the temperature in a club in order to make you buy more alcohol. Or even placing nuts on the bar to make you more thirsty. Fact is i might have been really thirsty and wanted another drinkwhich had nothing to do with the heat or what iv eaten
    good blog!

  5. Pingback: TA’S HOME WORK THIS WEEK :) « Leilla92

  6. I think it is important to stress the simple most important strength of correlation is the way that it makes it possible to identify the direction and strength of a relationship between 2 or more variables. It allows an investigator to measure relationships between variables with manipulation or control, which of course does not show cause but does help to give a general idea. Furthermore it is possible using correlations to predict the possible outcome/value of one variable if one knows the the value of the other. An example is in the grades attained by students and number of hours spent watching tv. You could predict that the compulsive tv viewer would achieve lower grades. However this does not prove that tv watching causes bad grades.

  7. It’s difficult to find the argument in your post as it’s pretty much a fact that correlation does not show cause and effect so therefore I think you could have considered slightly more about correlation as a whole in your blog entry.

    You used the example of more ice creams being sold on hot days; if the researcher wanted to look at what factors effect ice cream sales then observing this correlation would give them a good starting point for their research. As ‘secretdiaryofapsychstudent’ also pointed out it can additionally be a useful tool in predicting an outcome in certain situations.

    There are also many correlations which make no sense and better support your point that correlation does not show cause and effect simply because the two variables clearly have no effect on each other for example I recently read that people who use the internet explorer browser over other browsers such as chrome or safari have a lower IQ, clearly using internet explorer does not lower your IQ and using a browser like safari does not make it higher.

    While correlation is a useful tool in research, it is important to note how misleading it can be, I could, for example, publish a paper saying that more children develop an illness like meningitis than children who do not go to school; it may be true but that is almost definitely due to the fact more children go to school than don’t, it would certainly be harmful to suggest that school is directly linked to an illness which the correlation alone seems to imply.

  8. I think it’s more the fact of, yes there is a relationship between ice cream and hot weather but a correlation does not show which caused what? Ok in this example it may appear more logical that the hot weather caused the ice cream sales to increase; and not the increased ice cream sales causing the hot weather; but when considering other variables it may not be so clear cut. Also, there is no direct way of telling that one variable caused a direct change in another variable because there may have been a third, unaccounted for, variable that caused the change in both. Correlations are good to use as a starting point to establish whether there is an actual relationship and then an experiment can examine the variables further.

  9. Pingback: Comments for Julie « KLM's Blog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s