Class Notes (809,487)
COMM 2002 (65)
Lecture

# COMM November 4, 2013.docx Premium

4 Pages
95 Views

School
Carleton University
Department
Communication Studies
Course
COMM 2002
Professor
Heather Pyman
Semester
Fall

Description
Bivariate Analysis Univariate to Bivariate • Formulated a research question • Developed/discovered survey question that measure concepts of RQ • Stated H1 • Described each variable (recoded/missing values) o Distribution o Central Tendency o Dispersion • Develop a research Hypothesis (H1) to test relationship of independent to dependent variable • Need to be looking at distribution of variables before moving on to the next step • Get variables into proper shape and form from hypothesis that are stated • Recode variables -> want to have variables in shape and format that best suit the forms of hypothesis that need to be tested • If we have an interval level variable and wanted to look at the distribution, but it was hard to see it because of small percentage variables or have many different points • Decision on how to recode may depend on different thing -> recode it based on something that's already been seen • Maybe interested in ages -> recode variables into above or below • Or could recode the interval level variable by dividing it up into equal quartiles • To look at distribution, you recode in someway to look at it • How to look at a more manageable variable • Dichotomy of different things • When we have a nominal variable -> we recode based on some aspect of the responses that we have Look at some of the statistics that are associated with the responses • How responses vary around that value • Measures of central tendency • Interval level data, look at all three different central tendencies of data • Nominal -> Only mode • Standard deviation o If the mean is 10/20 o SD is 2 • 8-12 • 6-14 • 8-16 o SD is 1 • 9-11 o SD is 3 • 7-13 • Move to looking at how values of one variable are related to those of another • Independent variable and it's relationship to voting • The patter is exactly the same, first we look at the distributions, look for relationships and search for how they co-vary • If we look at gender, for example, are we more or less likely to vote? Bivariate analysis (analysis of two variables) • Never establish causality -> never say that one thing causes another • Explores relationships between variables • Search for co-variance and correlations • Cannot establish causality o Temporal order o Other variables o Explain behaviour by gender, marital status, etc. They are co-related with environmental behaviours • None are causally related o In attitudes and behaviour, we are never assuming that we are making those causal links because human behaviour is so diverse that we can never find one variable that explains the others Crosstabulation • Purpose o Connects the frequencies of two variables o Helps you identify any patters of association • Uses o Suitable for nominal and ordinal data o Difficult to look at something that is age and distribution Table Construction • Independent variable (IV) always goes in the column position • Dependent variable (DV) always goes in the row position • Column, row and total % all give different information • Column % reveal the relationship of the IV to the DV • Focus on only column positions -> for this class Example: HQ: Females are more likely to vote Males Females Total sample 100 200 300 33% 67% 100% Cross tab of gender vote Blank Males Females Vote Count
More Less

Related notes for COMM 2002

OR

Don't have an account?

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Join to view

OR

By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.