Focus Groups-Collective interview of a group of people, very focused.
-Systematic, methodological group session.
-Often used in needs assessments, program development, process evaluation (how the program
was received) and outcome evaluation (understanding the outcomes of the intervention).
-Must deal with ethical issues of informed consent, confidentiality, privacy (do not want too
much private/sensitive information given), and when dealing with stressful topics referring those
who need help.
-Range on a continuum of more structured (specific topic, questions and pre-determined what
issues will be focused on) and less structured (More exploratory, few pre-planned questions).
Strengths-excellent for exploratory research.
-Content and depth of answers, less superficial and can understand people’s reasons
-Can interact directly with participants and they can say things in their own words
-Allows participants/researchers to react and build on others responses.
-Can easily attain information from illiterate peoples or children.
Myths-Quick and cheap
-Requires professional moderators: may require them but not necessary, may just train someone
on the research team.
-Require special facilities: such as 2 way mirrors, not true.
-Must consist of strangers: most often they are not. Must be aware of colleague/friend
-will not work for sensitive topics.
-Produces conformity: actually quite the opposite.
-Something said must be validated by others: focus groups are a legitimate form of data
collection and thus can stand true on their own.
1) State the research purpose
2) Identify the moderator (select or train)
3) Develop an interview guide (set of questions with a sense of direction but can be revised
on the fly. Go from general to specific and use open ended questions, structured vs.
Unstructured and pay attention to the order of the questions. Have ice-breaker opening
questions, introductory questions (understanding of the topic but not the core questions),
key questions that you are most interested in, transition questions (move to a slight sub
topic), ending question (state a final position on the topic).
4) Recruit a sample-Use purposes, convenience samples and not random samples, decide of
a group composition (generally a homogenous group). Usually a group size of 7-10
individuals, consider the amount of time each person will get to talk. Consider the
diversity of the comments, want theoretical saturations (process of adding cases until
you’ve covered the full range of observations and no new info comes about.
5) Conduct focus groups-There are issues with self proclaimed experts, others interrupting
and friends participating.
6) Analyze and interpret data
Criteria of Objectives-Specific/Precise, measurable, purposeful standards with a realistic time
Internal Validity-The validity of inferences about whether the relationship between 2 variables
is causal. Concerned with research designs and deals with cause and effect relationships. Threats to Internal Validity-Any event that coincides with the independent variable and could
affect the dependent variable. E.g. effectiveness of a seatbelt campaign: increases in seatbelt
wearing could be a result of the program or the law, or fines or accidents in the community.
Maturation-Threat that some biological, psychological or emotional process within the person,
separate from the interventions, will change over time and change the results of tests.
Testing-Effects due to repeated testing of participants over time.
Instrumentation-Any change that occurs over time in measurement procedures or devices.
-e.g. blood pressure over time, pretest at the home, posttest at the clinic. See a change but it
could have been a stronger relationship had they used the same measure.
Statistical Regression-A problem of extreme values or a tendency for random errors to move
group results toward the average.
-Regression towards the mean, penny encouragement idea.
“Interrupted” time series with non-equivalent control group
Group A 01 02 03 04 X 05 06 07 08
Group B 01 02 03 04 05 06 07 08
-Control group must be as similar as possible to treatment to ensure the best results. Are the
results caused by the intervention or something else? Selection or history?
“Interrupted” time series with multiple replications
01 02 X 03 04 Remove intervention 05 06 X 07 08 remove intervention 09 010 X 011 012
Sometimes is unintentional.
Randomization-Randomly put people in groups (e.g. control groups vs. Experimental groups).
Design issue of internal validity.
Random selection-Pick a certain number of people for your group. Sample issue of external
1) Between participants (1 participant can only be in 1 group), randomized posttest only
Group A X1 01
Group B X2 02
-They will likely be equivalent at baseline but a limitation is that there is no pretest so we must
assume that they are equivalent we do not actually know.
-Because there is no baseline could have issues with the ceiling effect (top already all there is to
go is down and vice versa).
2) Between participants, matched, randomized, posttest only
Group A X1 01
Group B X2 02
-Match them on certain characteristics and then randomly assign. Able to detect small effects
3) Between participants, randomized, pretest, posttest
Group A 01 X1 02
Group B 03 X2 04 -Statistical comparisons can be on pretests and posttests, pretests and postests for each person
and compare group differences.
-There is no selection bia, statistical regression etc. all of the threats should be the same, and thus
do not matter. However, diffusion may be an issue.
4) Solomon 4 group design
Group A 01 X1 02
Group B 03 X2 04
R= Group C X1 05
Group D X2 06
-Allows for us to assess whether taking the pretest influences the posttest.
Can compare 01+03(should be the same, 02 +04 (should be different), 01+02 vs 03+04 (should
be different), 05+06 (should be different), 04+06 (should be similar), 02+06 (Should be
different), 02+05 (should be the same, critical reason for doing this study is to assess the effect of
5) Factorial Designs
Group A X1 Y1 01
Group B X1 Y2 02
R= Group C X2 Y1 03
Group D X2 Y2 04
-Looks at how many levels each factors has and creates all possible combinations. Is an
assessment of more than one independent variable.
-The main effect are the differences among groups for a single independent variable that are
significant, temporarily ignoring all other independent variables.
-Looks at the influence of one independent variable on another.
Experience Sampling-Studying experiences in naturally occurring contexts of everyday life.
Collect info about context and content of daily life of people.
-Written responses to open and closed ended questions at several random points throughout each
day of a normal week (signalling device prompts).
-Questions are about physical context, social content, activities, thoughts and feelings.
-Developed in early 1970’s by Csikzentmihalyi who did studies using pagers at random times.
Studied “flow” experiences where you are so absorbed in an activity at the time, place does not
impact you. Used diaries to find out enjoyable moments.
-Our subjective experiences are real, ESM allows for intrapersonal and interpersonal
comparisons. ESM samples experiences randomly.
-ESM is used in psychology of adolescents, experiences of work, work stress and work
satisfaction, marital satisfaction, parenthood and family relationships, experiences of depression,
substance abuse, eating disorder and exercise withdrawal.
-If ESM is appropriate the types of questions used are about groups of people or about situations
-they will be small purposive samples and will not include people with reading or writing
difficulties, hard of hearing or certain types of adult occupations.
-Can be with paper and pencil or PDA, must assess the duration, frequency and identify
parameters for the schedule e.g. set hours, personalized or randomly selected time in blocks. -Duration and frequency depend on the occurrence of the phenomena, how compliant
participants are, planned statistical analysis, degree of burden of filling out the form.
Participants in qualitative interviews are active participants
Most qualitative interviews are informal and nondirective and involves a mutual sharing of
experiences between the interviewer and interviewee. Leading questions and forcing answers
are avoided. The focus is on the interviewees perspectives and experiences.
Developing rapport is essential to qualitative interviews.
Qualitative interviews do not have a clear beginning or end and can be picked up at a later
time, the questions and order of questions are tailored to the person, interviewer shows
interest in responses and encourages elaboration, more like a friendly conversation, can occur
in a group setting, it is interspersed with jokes, asides, stores etc which are recorded, open
ended questions are common and probes are frequent, it is an interactive process, social
context of the interview is important and the interviewer adjusts to the interviewees language
Qualitative interviews differ from friendly conversations because they have a purpose, the
interviewer asks more questions and express more interest. It also includes repetition of
questions and themes.
Inductive approach to theorizing, usually no testing hypothesis.
Purposive and snowball sampling are often used because the topics studied are not usually
easily found with probability sampling. For example purposive sampling was used to find a
diverse group of exotic dancers, snowball sampling used to find street youth and mothers
using daycare in Quebec. Often qualitative researchers want to study hidden populations with
no sampling frame.
You should continue interviewing until you reach theoretical saturation but is often dictated
by the time and resources the researcher has available.
Incentives are given to participants as qualitative interviews are time consuming, must be
targeted to the specific population.
Research takes notes of gestalt (body language, relationship between interviewer and
interviewee and the context in which the interview takes place.
Selective transcription-interviewer feels it is unnecessary to transcribe the whole interview
and thus only transcribes the bits that are important to their researcher. However, fully
transcribed interviews are the best way to ensure findings are dependable and trustworthy.
An information is someone who is familiar with and engages in routine activities in the
culture of the interview, is currently involved in the culture (ex-members may also provide
useful insights), can spend time with the researcher and should be a nonanalytical individual.
Kvale created a typology of 9 different question types: introduction, follow-up, probing,
specifying, direct, indirect, structuring, interpreting and silence.
Follow up questions ask for expansion on a particular point, probing questions ask for
Direct questions are used towards the end of the interviewer to introduce specific topics that
have not come up yet.
Indirect questions are those the interviewer asks to get a sense of how the interviewee thinks
other people think, behave or feel. Structuring questions are those used to keep the interview on track.
An interview guide is a list of questions the researcher wants to cover, it does not have to be
followed chronologically, it is just a guide. It can be used towards the end of the interview to
make sure all the key topics have been covered.
Advantages: see the world from the perspective of the participant, lots of rich detailed data,
new theories can emerge due to the inductive approach.
Limitations: too much data can be overwhelming and time consuming, sample sizes are
usually small, however efforts can be made to make results generalizable.
The moderator of a focus group starts off the group, makes sure the group stays on tack,
makes sure a single person does not dominate the speaking and can effectively diffuse
In a typical focus group study there are 4-6 separate groups, however, time and money often
dictates the number of research groups.
Groupthink refers to a person’s natural desire to avoid conflict and lead towards a group
consensus even when they have differing opinions
In both qualitative and quantitative research they carefully examine empirical information
from social life, infer and make a conclusion. Both collect large amounts of fate, describe the
data and document how they collected and examined the evidence. Both also try to avoid
errors, false conclusions and misleading inferences.
Qualitative has many approaches to data analysis, quantitative uses hypothesis testing and
statistical methods. Quantitative gathers all data and then does data analysis whereas
qualitative starts data analysis during data collection (iterative). Quantitative tests hypotheses
whereas qualitative develops new concepts and theories. Quantitative uses numbers to
represent social life whereas qualitative uses words etc.
The goal of the researcher is to organize the details of the data into a coherent picture or
model and not necessarily prove cause and effect. Grounded theory.
For quantitative researchers conceptualization is done before data collection and analysis, for
qualitative it is done and refined throughout the data collection process and is grounded in
the data. Conceptualization is how the qualitative researcher makes sense of their data. They
often conceptualize, create concepts/themes, as they code data.
Coding in qualitative data is the process of organization raw data into conceptual categories
and creates themes or concepts. Coding is an integral part of the data analysis stage and in the
case of grounded theory moves the data toward theoretical generalizations.
Coding involves reducing the raw data into manageable chunks and creating analytical data
Open coding is the first look over recent data looking for critical terms, key events or themes
and noting them. It brings themes up from inside the data. There can be hundreds on initial
codes but the researcher usually writes about the ones with the most data.
Axial coding is the 2 look over data, starts with the initial open codes and focuses more on
these. Task here is to create linkages and themes between the open codes and how the open
codes may cluster together into themes. Often raises new questions, drops some codes and
analyses deeper other codes. Selective coding involves scanning all the data and previous codes and determining if there is
a core category that all data and codes fit into.
Analytical memo-discussion of thoughts and ideas about the coding process that a researcher
writes to themselves. Each concept has a memo. These memos draw the raw data into more