Flashcards in Research Methods Deck (85):
Three purposes of research
Explanatory: experiments, hypothesis testing.
Exploratory: to investigate new area
Descriptive: Describe the situation
Types of survey instrument
Questionnaires. Structured interviews
accurately informing your subject as to the nature of the research, information on questionnaire, covering letter or interview consent form, written consent to participate or verbal consent.
Inform subject about research, implied consent given by return of form, e.g. short postal questionnaire
All data treated as confidential, no information published will identify persons or organisations without their permission. Code umber included in right top corner of questionnaire
Not the same as confidentiality, can only be promised in some circumstances, no code numbers, no method of sending reminders.
A good response rate
Response rate =
(number completed/ number in sample) x 100
Easier to analyse, good if questionnaire is long, better if motivation is low, quick and easy to answer, creates false options.bias if not all options included, good design is vital, loss of spontaneity
Elicit a wide variety of responses, good for exploring, does not superimpose answers/expectations, difficult to summarise/analyse, response must be reported accurately, unpopular if used a lost in a postal questionnaire.
Avoid in questions
jargon/abbreviations, vague terms, leading questions, double-barrelled, double negative, presumptions, complex questions, boring questions, filter questions, social desirability effects, other peoples' influence.
Include in questions
Common concepts, simple and short questions
A 5 point scale, sometimes 3 point. Scales with no mid-point do not allow respondents to be noncommittal, can lead to missing responses. Mid point = neither agree nor disagree
pair of opposite statements
e.g. put a cross on line which corresponds with your pain (usually 10 cm)
Postal questionnaire design
Instructions, double sided (ensure that it is not cramped), good use of white space, avoid all capital letter, return date and address at the end, thank you at the end.
Instructions, single sided helps interviewer keep rapport
General questionnaire design
Indicate whether it's confidential/anonymous, sections, number all questions, demographic info at the end, avoid small type face, limited number of type faces, don't split question across pages, boxes or lines for open questions.
Increases response rate of postal questionnaires
Advance warning, letter of introduction, address envelope to respondent, incentive, stress confidentiality, make questionnaire look pleasant, do not make it longer than 8 sides, link questions and sections, include and return envelope and postage, contact phone number for queries, put a return date so that reminders can be sent.
Questionnaire covering letters
Aim of the questionnaire, stress confidentiality/anonymity, convey importance of study, encourage recipients to reply, signed by researchers and job titles, if sponsored mention this, used personalised approach, explain how and why sample was generated, reply by date included (2/3 weeks), no more than one side of A4
Questions to consider at pilot
How long did it take to complete, instructions clear, questions ambiguous, questions objectionable, layout clear and easy to follow, any topics omitted?
Quantitative research methods
self-administered questionnaire, structures or semi structures interviews, structures observation (non-participant)
Self administered questionnaire
A collection of questions administered and answered by respondents without the aid of an interviewer. E.g. postal questionnaire, e-mail surveys, web surveys, researcher handing out surveys to a class and collecting them and the end.
Advantages of self administered questionnaires
Low cost of data collection and processing, quick to administer, absence of interviewer bias, access to respondents who live at widely dispersed addresses or abroad, contact people who are rarely at home, convenient for respondents, used to screen a population for subsection who can then be interviewed.
Disadvantages of self administered questionnaire
need accurate and up to date list of population, low response rates, unsuitable for some respondents e.g. poor literacy, visually impaired, children. Cannot correct misunderstanding, no control over order of questions answered, do not know who answers questionnaire, cannot collect additional data, e.g. observation. greater risk of missing data, rely on respondent to complete questionnaire aided only by written instruction and covering letter
A research interview in which all respondents are asked exactly the same questions in the same order with the aid of a formal interview schedule.
The researched has a series of general questions that are in the form of an interview guide but is able to vary the sequence on the questions. The interveiwer usually has some latitude to ask further questions in response to what are seen as significant replies.
Advantages of interviews
Used for samples where there are no complete sits, interview schedules can be longer and more complete, obtain more detailed information, good response rates, used for people who are not literate, use prompts to explain what question means, interviewer has control over question order, useful for sensitive subjects, face to face allows collection of observational data.
Disadvantages of interviews
Expensive, difficult to use with diffuse samples, interviewer bias, acquiescence, social desirability, evaluation apprehension, respondents cannot consult records.
Where the interviewer influences the responses given, e.g. by how they are dressed or respond to comments
say yes to everything irrespective of what is asked
interviewee gives the socially acceptable answer rather than their own opinion
interviewee alters their views, beliefs or opinions because they want to give you what they think you need.
A technique in which the research employs explicitly formulated rules for the observation and recording of behaviour. These rules are articulated in an observation schedule.
The researcher observed but does not participate in the activities under investigation.
Advantages of structured/non participant observation
Allows the investigation of individuals' behaviour, identification, counting and characterising of events, more reliable and valid that self resorting.
Disadvantages of structured/non participant observation
people when the know they are being watched will alter their behaviour as a result of being observed. As an observation continues, the effect diminishes, the length of time does vary and depends on the type of observation.
User friendly data collection form clearly indicating what events/behaviours are to be observed, identification, counting and characterising of event, what events/behaviours are to be recorded. Ensure categories of behaviour/events do not overlap. Ensure it is easy for record behaviour/events on observation schedule.
Qualitative research methods
unstructured/ethnographic interviews, focus groups, observation, free response replies to interviews, free response replies to questionnaires
A guided conversation ether the interviewer has more power than the informant.
Advantages of unstructured interviews
Good for detailed information, people can expand on their opinion/behaviour, useful for sensitive subjects
Disadvantages of unstructured interviews
Bias, expensive and time consuming, hard to access subjects
A carefully planned series of discussions designed to obtain perceptions on a defines area of interest in a permissive, non-threatening environment.
Purpose is to understand how people feel or think about an issue, product or service. Involves 3-15 homogenic participants. Small groups: more details as everyone can contribute, discussion may dry up. Large groups: more interaction, group fragmentation and lack of contribution
Advantages of focus groups
Ideal for exploring factors that involve behaviour, natural social environment, quick results, cheaper than interviews, unanticipated themes may emerge.
Disadvantages of focus groups
Skilled moderator required, researched has less control, problems with group dynamics, groups are small so unrepresentative of general population, complex data.
Moderator of focus group
Welcomes participants to group, explains background to research, uses topic guide consisting of general open ended question to stimulate discussion. Uses specific questions once discussion develops. Group exercises to encourage interaction and explain views.
Assistant moderator of focus group
Takes care of recorders. Makes notes on the discussion and group dynamics.
Participants of focus group
Discuss their experiences and common on other's perspectives
Aim is to record in as much detail as possible the behaviour of participants with the aim of developing a narrative account of that behaviour.
Advantages of unstructured observation
Does not rely on reports by others
Disadvantages of unstructured observation
Not possible for large samples, Hawthorne effect, triangulate to determine observer effect, labour intensive
Validity of qualitative research
Inherently high. Reduced by preconceptions of researcher or external influences. Improve through triangulation (use different data collection techniques to collect the same data)
Reliability of qualitative research
Low. Data is context specific - may get different results every time you visit people.
Generalisabilty of quantitative data.
Low. Aim of data is to explore not generalise.
Quantitative data processing
Consists of coding data using a developed coding framework
A specification of codes for all variables (e.g. questions in a questionnaire) and values (i.e. potential responses to) these variables in the dataset.
Reason for numerical coding
Statistical databases use numerical codes to analyse data.
Coding a data collection form
a uniques code should be assigned to each respondent or case for identification.
Coding closed questions
Usually have a limited number of responses. Easily coded. Muse include code for missing data.
E.g. male = 1, female = 2, not completed = 99.
Coding multi-response questions
Some questions may require the respondent to provide more than one response. For these questions, each possible response may need to be coded as if a separate variable. E.g. 1a, b, c, d, e, f all need to be coded
Coding open questions
Respondents can answer these questions in any way they want, frame is based on the actual responses by listing different responses, grouping them and assigning codes to the identified groups.
Qualitative data analysis
Process where effort is used to identify themes and construct hypothesis from the data.
Process of qualitative data analysis
Organising noticing themes, common features with regard to a topic, marking segments of text to show data which cover the same topic and note when they are in original text, linking together text segments with common themes (=coding), de-contextualisation (taking text segments out of original context), re-contextualisation (putting data of a common topic together).\
A computer package for coding qualitative data
Manual method of coding qualitative data
Cut up and put into a folder method
The attempt to derive generalisable new knowledge by addressing clearly defined questions with systematic and rigorous methods.
hypothesis testing, e.g. evaluating/comparing interventions
Quality improvement process that seeks to improve patient care and outcomes through systematic review of care against explicit criteria and the implementation of change.
A set of procedures to judge a pilot's merit by providing a systematic assessment of its aims, objectives, outputs, outcomes and costs.
Mesures current service without reference to a standard. Does NOT require REC review
Experimental research design
Treatment vs placebo
cross-sectional research design
data collected from sample o a single occasion
Longitudinal research design
Data collected from a sample on more than one occasion
Case study research design
Detailes and extensive analysis of a case
Comparative research design
Comparison of two or more cases
Researcher collects their own data
Analyse data obtained by someone else
Sample size calculations. Random sample - each person has an equal chance of being chosen.
No sample size calculation.
Convenience - ask people because they are easily accessible/ you know them
Purposive - you select a group of peoples on purpose because they have a particular characteristic of interest.
Snowball - recruiting individuals and asking them if they know anyone else who would like to participate, and repeat.
Get the same result every time
Is the group being studied typical of others in the same situation
Does you data collection measure what it is intended to measure?
Are the research findings generalisable to the wider population of interest?