How Much Do Polls Tell Us?

Journalists seem to be doing their best to bash polling, as indicated by this not so insightful piece from the Atlantic.  Having said that, there seem to be a lot of naive beliefs about both polling and politics that are commonplace, and in that sense, this may be exactly the sort of problem one’d run into in extrapolating from the polling data.

The real problem with polling is that every poll is true but limited and biased.  Yes, we asked real people and they gave real answers–which hopefully are honest (but may not be).  But we know that we only caught a small slice of the “reality.”  The real challenge of polling is not always (or even necessarily) to find means of conducting “better” polls, but to find better means of extrapolating based on our understanding of politics.

So, suppose you find that your poll of “Hispanics” include some proportions of Puerto Ricans, Cubans, and Mexicans in a given area…as well as across the entire country.  Are all these “Hispanics” more alike within this collective category than between themselves?  That is, do Mexicans have more in common with Puerto Ricans than they do with the average voter of all race and ethnic groups?  Do different ethnic groups in the same locality have more in common than with the members of the same ethnic groups in different localities?  If they are more alike within the ethnicities than within the same locality, perhaps you should use data from the national pool to extrapolate from the data you have in the locality.  Do young Hispanics think like younger people of other ethnicities, or is there something different that makes the young Hispanics different?  Again, the same logic applies.

The process of drawing inferences from polls requires a good deal of thinking intermixed with data work.  You cannot arbitrarily assert that all Cubans, across all localities, are alike and extrapolate from this assumption/assertion:  you want to see if this is indeed true in the data you have first.  But with some careful thinking, close attention to the data, creativity, and better understanding of the way political winds are blowing, there is plenty that one can draw out from a poll–which many are willfully refusing to do on the account of closedmindedness.


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