Everything I Needed to Know About Life I Learned from Statistics........
Well, maybe not everything, but a lot! As I plough through my grad stats course, I feel myself continually needing to connect the practical, concrete concepts I'm learning to broader, more abstract social phenomena. In thinking about this, I've concluded that there are many aspects of statistical analysis that may/can be considered analogous to social life:
1) Relationships are like Pearson's Correlation Coefficients: And I'm not just talking about romantic relationships. ALL relationships may be compared to this magical little statistical measure. You see, the Pearson's 'r' measures the degree of association between two variables, ranging in value from zero to negative or positive 1; a relationship of 1 indicates perfect association between two variables, and zero indicates that there is no assoication. Doesn't this remind you of people you know? There are those rare, beautiful ones that you seem to just connect with so well; you're postive about things that they're positive about, and negative about things that they are as well. If you have strong thoughts on a certain issue, you're almost certain that they'll feel the exact same way. Alternatively, there are those people that I like to personally term the 'Anti-Me.' You know who I'm talking about: those annoying philistines who are your moral juxtaposition, like people who listen to Kevin Federline music because they thinks it's good, not because it's funny. Like a Pearson's r value, your measure of association with these people is at a value of zero--the two of you really have no association. And, finally, there are those ambiguous in-between values that statisticians are still unsure about, asking themselves, "does this equate to a 'moderate' relationship? A weak one?" while continually scratching their heads. These are the people who you occasionally have a fabulous time with, yet they annoy the fuck out of you at the next event. So, the lesson is: Don't get mad about this lack of association. Instead, embrace your perfect 1's, and recognize that the only way you can transform a variable is in a strictly artifical sense that does not accurately reflect reality.
2) Perfect Models Don't Exist in Real Life: As a social scientist rather than a natural scientist, I study statistical data that is based on actual human experience, rather than abstract mathematical concepts. In this, we generate predicitive equations; these equations are the 'models' for our data, which offer the best predicitions of the dependent variable given our set of independent variables. However, our results NEVER perfectly match these models. Rather, they constantly deviate from the model we've predicted, thus making our model a nice idea, but not an actual reality. Thus, I've learned that 'model' anything is just that--a model. So, while we may point to individuals or groups, calling them the 'model family', 'model couple', 'model student' or 'fashion model', I like to take comfort in knowing that they, too, have error terms stamped all over their perfect little facades.
3) People are Like Data Sets: When you work with data sets, you're bound to come across two potential types of errors: random errors and systematic errors. Random errors can be produced by little human mistakes in data entry or question asking during a survey, and pose no serious threat to the end results of your analysis. Alternatively, systematic errors bias your entire analysis. Lesson learned: People are totally like data sets! Many, or probably ALL have random 'errors', or little annoying qualities to them that won't horribly impact the entire product of their life and/or relationships. However, we do come across the occasional douche-bag of a human being who's general stance towards everyone and everything will destroy all that is in her/his wake.
and last, but certainly not least:
4) You Can Only Get So Far With Quantitative Measures: There are some things happening in the social world that can simply not be accurately studied/analyzed using numbers, and thus, qualitative research methods have to be employed. Analogously, in life, we cannot measure our total experience using income, value of the home we live in, the vehicles (or, in the case of Calgary, SUV's) we drive, or any other tangible product that we can know the quantitative value of.
1) Relationships are like Pearson's Correlation Coefficients: And I'm not just talking about romantic relationships. ALL relationships may be compared to this magical little statistical measure. You see, the Pearson's 'r' measures the degree of association between two variables, ranging in value from zero to negative or positive 1; a relationship of 1 indicates perfect association between two variables, and zero indicates that there is no assoication. Doesn't this remind you of people you know? There are those rare, beautiful ones that you seem to just connect with so well; you're postive about things that they're positive about, and negative about things that they are as well. If you have strong thoughts on a certain issue, you're almost certain that they'll feel the exact same way. Alternatively, there are those people that I like to personally term the 'Anti-Me.' You know who I'm talking about: those annoying philistines who are your moral juxtaposition, like people who listen to Kevin Federline music because they thinks it's good, not because it's funny. Like a Pearson's r value, your measure of association with these people is at a value of zero--the two of you really have no association. And, finally, there are those ambiguous in-between values that statisticians are still unsure about, asking themselves, "does this equate to a 'moderate' relationship? A weak one?" while continually scratching their heads. These are the people who you occasionally have a fabulous time with, yet they annoy the fuck out of you at the next event. So, the lesson is: Don't get mad about this lack of association. Instead, embrace your perfect 1's, and recognize that the only way you can transform a variable is in a strictly artifical sense that does not accurately reflect reality.
2) Perfect Models Don't Exist in Real Life: As a social scientist rather than a natural scientist, I study statistical data that is based on actual human experience, rather than abstract mathematical concepts. In this, we generate predicitive equations; these equations are the 'models' for our data, which offer the best predicitions of the dependent variable given our set of independent variables. However, our results NEVER perfectly match these models. Rather, they constantly deviate from the model we've predicted, thus making our model a nice idea, but not an actual reality. Thus, I've learned that 'model' anything is just that--a model. So, while we may point to individuals or groups, calling them the 'model family', 'model couple', 'model student' or 'fashion model', I like to take comfort in knowing that they, too, have error terms stamped all over their perfect little facades.
3) People are Like Data Sets: When you work with data sets, you're bound to come across two potential types of errors: random errors and systematic errors. Random errors can be produced by little human mistakes in data entry or question asking during a survey, and pose no serious threat to the end results of your analysis. Alternatively, systematic errors bias your entire analysis. Lesson learned: People are totally like data sets! Many, or probably ALL have random 'errors', or little annoying qualities to them that won't horribly impact the entire product of their life and/or relationships. However, we do come across the occasional douche-bag of a human being who's general stance towards everyone and everything will destroy all that is in her/his wake.
and last, but certainly not least:
4) You Can Only Get So Far With Quantitative Measures: There are some things happening in the social world that can simply not be accurately studied/analyzed using numbers, and thus, qualitative research methods have to be employed. Analogously, in life, we cannot measure our total experience using income, value of the home we live in, the vehicles (or, in the case of Calgary, SUV's) we drive, or any other tangible product that we can know the quantitative value of.
posted by Soci Womyn at 1:33 PM
0 Comments:
Post a Comment
<< Home