Qualitative vs. Quantitative methods
The miner and the traveler
• Miner • Traveler
– Knowledge is buried – We create knowledge
metal – Postmodern
– Knowledge is given constructive
– Knowledge exists understanding
– Knowledge is waiting
to be uncovered,
uncontaminated by the
• Silverman’s points
• Differences between qualitative and quantitative
• Different methods for different purposes
• A combination perspective
• An example of a quantitative study “Effects of
TV advertising on Chinese consumers”
Silverman p. 31, 2001
• ”I conclude […] by observing that an
insistence that any research worth its salt
should follow a purely quantitative logic
would simply rule out the study of many
interesting phenomena relating to what
people actually do in their day-to-day
Problems with qualitative studies
according to Silverman
• Lack of theoretical framework and too
little theory building from the data – just
• Too much ”putting one self in the
• 1. Epistemological contradictions
– Quantative methods. Developed with regard to objective, precise,
correct and covering depiction of a reality and grounded in
– Qualitative methods. Developed to understand people’s
construction of an image of society and their interpretation of the
social relations they are a part of, and they are grounded in
constructivism or idealism.
• 2. Subject/object
– Fundamental contradictions in the perceptions of human beings
– Qualitative methods are often based on a perception of human
beings as active agents, who through their consciousness of social
relations take part in the creation and recreation and development
of a society.
Qualitative or quantitative
1. The topic of the survey (the nature of the
subject under investigation)
2. The way we look at the subject under
investigation (the problem)
3. The purpose: what kind of knowledge do
we want to generate?
4. Time and ressources
Data Calculations of Identification of
Analysis Argumentation is Reasoning and
based on numbers and argumentation that
on systematic, are not based
statistical relations simply on
between the numbers statistical relations
regularites in the way Riddle-solving
different variables are
associated with each
Quantitative methods Qualitative methods
Research design: Low High
Standardisation and High Low
Production of Fragmented Holistic
Assumptions on Stable and static Unstable and dynamic
coherence in real
Selection of Representative Who has most knowledge
respondents and information
Expected use of Formulation stage: big Formulation stage: small
time Analysis stage: small Analysis stage: big
Source: Table collected by Susanne Jensen, AAU
Preparation of a questionnaire
• Random sample or total population?
• Are the respondents in the sample representative
for all relevant units?
• Necessary sample size: Rules of thumb
– Smallest subgroup analysed in the sample must not be
smaller than 35-50 units
– If the smallest subgroup amounts to 5% of the
population the sample must include at least 1.000 units
Necessary sample size
• The necessary sample size depends on the
size of proportion p in the population. P is
the share of the population that for instance
vote on a particular party or who has a
particular attitude to the environment
Proportion p Necessary sample size
0,40 or 0,60 50
0,30 or 0,70 80
0,20 or 0,80 200
0,10 or 0,90 600
0,05 or 0,95 1400
Areas to be careful about when doing a questionnaire, where
the reliability of the study risks being reduced
• Sending out the questionnaire
– Are the respondents sampled in an appropriate way?
• Do all relevant persons receive the questionnaire?
– Which situation is the respondent in?
– Are the intentions with the questions understood?
– Are the answers placed correctly in the questionnaire?
– Are the possibilities to answer exhaustive?
– How many respondents answered?
– Persons and questions not answered: any pattern?
• Coding the data
– Is it done precisely?
– Is it the relevant categorizing?
• Presentation of the results
– Is the data over-interpreted?
(source: Ib Andersen, Den skinbarlige virkelighed, 1999 and
Susanne Jensen, AAU)
Types of quantitative analysis
• Univariate analysis: analysis of a variable and its
distribution on the units of investigations
• Bi-variate analysis: analysis of two variables and their
mutual co-variation. The co-variation may be investigated
by cross tabulation the two variables. The results is
shown in a table of cross tabulation
• Multivariate analysis: analysis of the co-variation of more
than two variables. Typically it is used to make probable
one or several causal connections and to put forward a
Look at Tai and Pae’ text
• Table 1, p.60 – what does it show?
• From either quantitative or qualitative
• To a combination….
Mix of methods
Can be used to
• supplement and
Ex. Reddy concludes on the Danes from one
village, Gundelach cannot confirm hypotheses
about changes in the Danes’ values regarding
materialism through questionnaires. A mix would
have strenghtened both studies.
• You’ve been asked to study Danish and
German families regarding their holidays
by the Danish Tourism Board.
• Please consider how quantitative and/or
qualitative methods can be applied, alone
or as a mix of the two methods.