Quantitative data analysis Tuesday, October 23, 2018 3:28 PM Constructing contingency tables on final* Q: during an interview, what should an interviewer do when a respondent provides an incomplete or inappropriate response? A: probe a little further Quantitative Analysis · Converting all our data to numbers · Coding- assign numbers to our variables · Ex- variable religion- arbitrarily give our attributes numbers. Christianity- 0, Buddhism- 1, Catholicism- 2, etc. doesn't matter, as long as you have a record of what the numbers mean Must be mutually exclusive and jointly exhaustive- same characteristics our variables carry . After coded and put in software- data set ttset of all your codes, all numbers) . Code book is what tells us what the numbers mean. Variable name and that the numbers represent · Stats Canada has data bank of questions and codes, can use their attributes and codes ttor arbitrarily come up with your own code) . Code book- guide, key Missing data- not stated, don't know, other, don't leave anything blank. Code everything even our missing values Quantifying Data Data cleaning- 2 things to clean before analyze our data- . First- clean for mistakes ttex. Face to face survey, writing down responses and they already know the codes, but wrote down the wrong code) put as don't know, other, or something if you cant figure it out · Second- contingency ttex. Question how many babies have you given birth to and the male person says 6, there's a mistake so we have to correct. Either they shouldn't have answered the question, or they answered their gender incorrectly earlier in the questions) we either don't include in data set because we cant solve problem, or we can code as something else. If everyone is making same mistake that might be interesting itself · Once clean, coded, start process of analysis · ID number is treated as a variable, individuals treated as attributes Q: What term refers to the process of reducing a wide variety of idiosyncratic items of info to a more limited set of attributes composing a variable? A: Coding ttreducing it to a more limited set of attributes) Ex: variable- country - state country you are born in, maybe the answers will be too much, so we will convert the variable to one called continent ttso we don't have so many attributes, to a more limited set) Univariable Analysis · Descriptive- cannot explain or explore anything with one variable
· Most frequent way is through frequency distribution- how our attributes are distributed- how many are Christian, Buddhist, etc. how many people selected each of our attributes. How frequently are our attributes selected. Constructing a frequency distribution table · First thing you do is LABEL the table ex: Table Showing Frequency Distribution of Soci 217 Midterm Grades · One column attributes ttgrades in this case) other is for frequency, other is percentage . Describe our variable in terms of percentages ex- 17% of the class got 60% Central Tendency ttmeasures of central