Ways of Looking Study Design by Kumar (Quantitative (Prospective study…
Ways of Looking
Disadvantage: less specific, less precise, not the same structural depth. Distinction between study design and method of data collection less clear.
Advantage: Main focus is to understand, explain, explore, discover
and clarify situations, feelings, perceptions, attitudes, values, beliefs, and experience of a group of people
Study based deductive argument, flexible and emergent in nature
Often non-linear, non sequential
Study with a selection of people, parameters of scope and information gathering, method and process flexible and evolving.
Adherence to concept of respondent concordance, whereby you as a researcher make every effort to seek agreement of your respondents, within your interpretation
power gap between researcher and study population is far smaller
replication of study almost impossible
Tricky for researcher bias in all qualitative study
Qualitative research recommended for
exploring values, beliefs, perceptions & meanings
Study design for qualitative research: Case Study
Individual, group, A community. Treat the study population as one entity.
Single method or multiple methods: interview, secondary sources, group interviews. etc.
Oral history - "approach used to study perceptions, experiences and accounts of an event or gathering historical knowledge as viewed by individuals."
Gathering people's thoughts in their own words about an event, topic etc.
Recording peoples opinion or interpretation, based on their own experience or passed from other sources.
"Memory is the core of oral history" (Ritchie 2003)
Focus Groups / Group interviews
Researcher raises questions to stimulate discussion
8-10 people max - hand picked by researcher based on their background or expertise
Records of the discussions becomes basis for the research
Perfect to explore diversity in opinions on different issues. Not ideal to find out the extent of the diversity. challenges -
some people might dominate conversation
Participant observation: observation of the real life situation, activities, researcher gets involved in the activities of the group, create a rapport with group members.
sufficient time spent with the group to gain much deeper and accurate information but possible researcher bias.
Community discussion forums
designed to collect opinions, attitudes of the community with regard to the community issues ad problems.
Most common approach, can be specific, well structured
validity, reliability and can be explicitly recognised and defined.
respondent accordance not relevant
Study can be replicated for verification and reassurance
Findings can be replicated and tested.
Classification of studies
1) Number of contacts within study population
2) the reference
period of the study
3) The nature of the investigation
Cross sectional study design
- most common study
select a study population and find out the required information. Example: "the health needs of A community"
The before and after study
measure impact by comparing before and after certain intervention and assess difference. Some use one control group, others use two control groups. (some issues - some people move on, you cant track people) Example: Impact of advice from GP in the health of A community
issues: regression effect: research instrument ie questionnaire makes people think about their attitude at the time of the post-test.
- Measure extent of change in a phenomenon situation, identify patterns. Main issues: conditioning effect = when people get contacted all the time and they lose interest, information not accurate sometimes
Retrospective study design - assess what happened in the past. Example the health status of A community in the 1920s
Prospective study design - check outcome of the future
Example: To determine the health status of A community
The retrospective-prospective study - focus on past trends and study it into the future. Example: the impact of maternal services on the health of babies within A community
Nature of the investigation - experimental, non-experimental and quasi-experimental
experimental - look at the effects and attempt to find the cause
study in 'controlled' - ie in a room / lab or 'natural environment', random group or particular group.
non-experimental - observe the cause and effect and assume both to be linked? random or particular group
semi-experimental - a bit of both.
compare effectiveness of different treatment modalities
experiment or non experiment
sometimes also include a placebo sesign
Cross over comparative experimental design
interventions crossed over - experimental control becomes control and vice versa.
forecasting by extrapolating present and past trends.
Example Changes in political preference of a study population over x period of time
existence of a common characteristic such as year of birth, graduation or marriage. Example: Identify employment pattern of all people in the A community who graduated in 1930s.
similar to trend and cohort.
Select a few families for example and find out how much they spend on household items in A community.
Check their expenditure over a period of time to identify a pattern.
Blind study fake drug
Double blind study: eliminate researcher bias by concealing the identity of the experimental and placebo groups from the researcher. So neither researcher or participant know who is receiving the real and the fake drugs.
Quantitative and Qualitative combo
Combination of both is more comprehensive
Start with qualitative to determine the spread of diversity
then move to quantitative to quantify the spread
then back to qualitative to assess the spread and explain observed patterns
Reflective journal log
can be used in combination
with any other method.
Collation of researcher's thoughts