Chapter 15 Social Research

Quantitative and Qualitative Data Analysis

Chapter 15

Introduction

Quantitative or Qualitative? What is the difference been qualitative and

quantitative? The distinction between qualitative and

quantitative data is not as important as the distinction between the strategies driving their collection

Introduction

Quantitative data analysis Analysis that tends to be based on the

statistical summary of data Quantitative researchers typically focus on the

relationship between or among variables, with a natural science-like view of social science in the backs of their minds.

Introduction

Qualitative data analysis Analysis that tends to results in the

interpretation of action or representations of meanings in the researcher's own words

Empathic understanding or an in-depth, thick description

Quantitative Data Analysis

Presumes one has collected data about a reasonably large, and sometimes representative, group of subjects, whether these subjects are individuals, groups, organizations, social artifacts, etc.

The data does not always come in the form of numerical data


Sources of Data for Quantitative Analysis When data is collected by researcher, coding

is an important first step Coding is the process by which raw data are

given a standardized form. This means making data computer usable.

For example, if you are coding gender – you may have Male = 1 and Female = 2

The assignment of numbers to words is arbitrary


Elementary Quantitative Analyses Descriptive statistics

Statistics used to describe and interpret sample data

Example Fifty-five percent of the people sampled were

married.


Elementary Quantitative Analyses Inferential statistics

Statistics used to make inferences about the population from which the sample was drawn

Example Men are significantly more likely than women to

have been employed full-time.


Univariate analyses Analyses that tell us something about one

variable


Bivariate analyses Analyses that focus on the association

between two variables


Multivariate analyses Analyses that permit researchers to examine

the relationship between variables while investigating the role of other variables

Univariate Analysis

Measures of Central Tendency Mode

The measure of central tendency designed for nominal level variables. The value or category that occurs most frequently. It can be computed for any variable because all ordinal and interval level variables are also nominal.

Univariate Analysis

Measures of Central Tendency Median

The measure of central tendency designed for ordinal level variables. The middle value when all values are arranged in order. Can also be used for interval variables because they are also ordinal variables.

Univariate Analysis

Measures of Central Tendency Mean

The measure of central tendency designed for interval level variables. The sum of all values divided by the number of values.

Univariate Analysis

How does a researcher know which measure of central tendency (mode, median, or mean) to use to describe a given variable? Do not use a measurement that is

inappropriate for a given level of measurement Example: Mean or Median for a nominal level

variable like gender

Univariate Analysis

Variation Frequency Distribution

A way of showing that number of times each category of a variable occurs in a sample

Assume we have 20 people in our sample, with 17 females and 3 males

Frequency Distribution

GENDER FREQUENCY %

Female 17 85

Male 3 15

Total N = 20 100

Univariate Analysis

Variation Examining frequency distribution, and their

percentage distribution is a good way of understanding variation in nominal or ordinal variables

Example If you are looking at gender and discern that

100% of your sample is female and 0% is male, you know that there is no variation in gender in your sample.

Univariate Analyses

Measures of Dispersion of Variation for Interval Scale Variables

Measures of dispersion Measures that provide a sense of how spread

out cases are over categories of a variable

Univariate Analyses

Measures of Dispersion of Variation for Interval Scale Variables Range

A measure of dispersion or spread designed for interval-level variables. The difference between the highest and lowest values.

Univariate Analyses

Standard Deviation A measure of dispersion designed for interval-

level variables and that accounts for every value's distance from the sample mean

The standard deviation has properties that make it useful in measuring variation when the variable is normally distributed

Univariate Analyses

The graph of a normal distribution is bell-shaped and symmetric

In a normal distribution 68% of cases would fall between one standard deviation above the mean and one standard deviation below the mean

Standard deviation is not as useful if the variable is not normally distributed.

Bivariate Analyses

Examining the relationship between variables Crosstabulation is the process of making a

bivariate table to examine a relationship between two variables

Bivariate Analyses

Measures of association Measures that give a sense of the strength of

a relationship between two variable – or how strongly two variables “go together”

Bivariate Analyses

Measures of correlation Measures that provide a sense not only of the

strength of the relationship between two variables, but also the direction of the association

Pearson’s r is a measure of correlation designed for examining relationships between interval level variables.

Stop and Think

Would you expect the association between education and income for adults in the US to be positively or negatively correlated?

Bivariate Analyses

Inferential Statistics P-value

Allows the reader to make an inference about the relationship between variables.

The typical cut off is 0.05, p<.05

Multivariate Analysis and the Elaboration Model Why would a researcher want to examine

more than two variables at a time?

Multivariate Analysis and the Elaboration Model Elaboration

The process of examining the relationship between two variables by introducing the control for another variable or variables

Multivariate Analysis and the Elaboration Model Control variable

A variable that is held constant to examine the relationship between two other variables

Multivariate Analysis and the Elaboration Model Partial relationship

The relationship between an independent and a dependent variable for that part of a sample defined by one category of a control variable

Multivariate Analysis and the Elaboration Model Four kinds of elaboration

1. Replication

2. Explanation

3. Specification

4. Interpretation

Multivariate Analysis and the Elaboration Model Replication

A kind of elaboration in which the original relationship is replicated by all of the partial relationships

Multivariate Analysis and the Elaboration Model Explanation

A kind of elaboration in which the original relationship is explained away as spurious by a control for an antecedent variable

Multivariate Analysis and the Elaboration Model Specification

A kind of elaboration that permits the researcher to specify conditions under which the original relationship is particularly strong or weak

Multivariate Analysis and the Elaboration Model Interpretation

A kind of elaboration that provides an idea of the reasons why an original relationship exist without challenging the belief that the original relationship is causal.

Qualitative Data Analysis

The outputs of qualitative data analyses are usually words, the inputs are also usually words – typically in the form of extended texts

Data is almost always derived from what the researcher has observed, heard in interviews, or found in documents


Social anthropological versus interpretivist approaches Social anthropologists (and others, like

grounded theorists and life historians) believe that there exist behavioral regularities (for example, rules, rituals, relationships, and so on) that affect everyday life and that it should be the goal of researchers to uncover and explain those regularities.


Social anthropological versus interpretivist approaches Interpretivists (including phenomenologists

and symbolic interactionists) believe that actors, including researchers themselves, are forever interpreting situations, and that these, often quite unpredictable, interpretations largely affect what goes on.


Does qualitative data analysis emerge from or generate the data collected? The question of which comes first

Data or ideas about data


The strengths and weaknesses of qualitative data analysis revisited Strengths

Can produce theories More likely to be grounded in the immediate

experiences of those participants than in the speculations of researchers.


The strengths and weaknesses of qualitative data analysis revisited Weaknesses

Generalizability


Are there predictable steps in qualitative data analysis? First researchers code their own data or

acquire computer-ready data Other steps are much more fluid Typical flow includes data collection –data

reduction—data displaying—conclusion drawing and verification


Data Collection and Transcription Several software packages exist to facilitate

the processing of qualitative data Qualitative data software packages have

many pros an cons and should be considered carefully before adopting.


Data Reduction The various ways in which a researcher orders

collected and transcribed data Coding and memoing are common data

reduction techniques


Coding The process of assigning observations, or data,

to categories In qualitative analysis, coding is more open-

ended because both the relevant variables and their significant categories are apt to remain in question longer


Coding The goal of coding is to create categories that

can be used to organize information about different cases

Assigning a code to a piece of data is the first step in coding

The second step is putting the coded data together with other data coded the same way


Coding Types of Coding

One purpose of coding is to keep facts straight – called descriptive coding

Coding to advance your analysis is analytical coding

The preliminary phase of analytical coding is called initial coding

Initial coding eventually becomes focused coding, which is concentrating or elaborating on codes specific to analysis


Coding Memos

Extended notes that the researcher writes to help herself or himself understand the meaning of codes


Data displays Visual images that summarize information

Summary

Quantitative data analyses Qualitative data analyses

Quiz – Question 1

Measures of central tendency do not includea. the mode.

b. median.

c. mean.

d. standard deviation.

Quiz – Question 2

In a frequency distribution, we area. displaying the number of cases that fall in

categories.

b. showing the connections between descriptive statistics.

c. examining the central tendencies of variables.

d. testing out our coding schemes.

Quiz – Question 3

As a measure of dispersion, a _______ tells us how far the mean is from individual scores.

a. range

b. standard deviation

c. mode

d. regular distribution

Chapter 15 Social Research

Documents

quantitative analysis

interval variables

sample data example

raw data

interval level variables

intervallevel variables

nominal level variables

ordinal level variables