test 2 Flashcards
what are the four main sources of topics?
- theory
- personal experience
- replication
- literature and previous studies
what is a theory?
organised body of concepts/generalisation/principles subjected to investigation
what are three problems with broad topics?
- enlarges scope beyond reason
- complicates organisation of review
- too general, difficult to carry out/interpret
what are some suggestions that can narrow topics?
talking to experts and read secondary sources that provide overviews
how do quant and qual studies differ when the topic is narrowed?
quant tend to narrow topic initially and qual narrow throughout
what are some factors that make a topic researchable?
- can be investigated throughout collection and analysis of data
- theoretical or practical
- contribute to educational process
- ethical
what are some factors that make a topic non-researchable?
-matters of opinion
what does a formal statement of quant research topic involve?
- identifies variable of interest
- describes specific relationship between variables
- identifies nature of participants
what are some functions of a literature review?
- determine what has been done
- provide insight to develop framework into which topic fits
- identifies useful methodological strategies
- facilitates interpretation of results
what is a literature review?
involves systemic identification, location and analysis of documents containing info related to research problem for question
what are some general recommendations for the scope of a literature review?
- bigger isn’t better
- heavily researched topics provide references to focus only o major studies
- lesser researcher require reviewing any study related in meaningful way
what are the four stages when conducting a lit review?
- identify keywords
- identify sources
- abstracting info founding references
- analysing, organising, and reporting lit
what are the seven steps of locating, reviewing, summarising, and classifying references in lit reviews?
- read abstract
- skim article
- record complete bibliographic info
- classify/code article
- summarise
- identify thoughts about what is important
- indicate direct quotes properly
what are the characteristics of sources in lit reviews?
- primary and secondary
- empirical and opinion
what are the two characteristics of meta-analysis?
- review as inclusive as possible
- results of each study are translated into stat called effect size
what is an effect size in a meta-analysis?
difference between means for experimental and control in CG standard deviation units
what does the average of all effect sizes summarising in meta-analysis?
summarises overall effect of studies
what is an inductive view of hypotheses?
generalisation made from number of observations (qual)
what is a deductive view of hypotheses?
derived from theory aim at providing evidence to support/expand/contradict (quant)
what is a good hypothesis?
- based on reasoning
- provides reasonable explanation for predicted outcome
- clearly states expected relationships between variables
- testable
what is a non-directional type of quant hypothesis?
statement that no relationship or difference exists between variables
what is a directional type of quant hypothesis?
statement of expected direction of relationship/difference between variables
what is a null type of quant hypothesis?
stat statement that no statistically significant relationship/difference exists between variables
how are hypothesis tested?
- stat analysis of data
- importance of results regardless of outcome
- results support or fail to support hypotheses, never disprove or prove
what does writing a lit review involve?
- report primarily on original sources of scholarship
- gives justification
- provide background
- describes, summarises, evaluates, compares and integrates content of primary reports
what are some reasons for lit reviews?
- identify gaps
- avoid repeating mistakes
- build on other’s work
- increase breadth of knowledge
- identify people in same field
what is a systemic review?
lit review focused on single question which tries to identify, appraise, and synthesise high quality evidence relevant to question
what is a meta-analysis?
statistical technique often used in systemic reviews enabling results from no.of quant studies combined into common metric to determine average effect of technique
what is an integrative review?
lit review that attempts to combine experimental and non-experimental studies
what can integrative reviews be used for?
review theories, evidence, or to analyse methodological issues
what is a qualitative review?
methods for combining qual studies, techniques include meta-synthesis, formal grounded theory and meta-ethnography
why do we use systemic reviews?
- minimise impact of bias/errors
- end confusion
- highlight findings from different studies
- can mitigate need for further trials
what is the significance of systemic reviews?
large amount of medical lit requires clinicians/researcher are likely to rely on one systemic review to make informed decisions
why are systemic reviews necessary?
volume of material makes it impractical for clinician to remain up to date
what are the benefits of systemic reviews?
- detect small but significant treatment effects
- save time/money
- clear method and results
- define boundary of what is known and not
who can undertake systemic reviews?
- cochrane/campbell collabs
- regulatory bodies
- academics
- health tech assessment
- multidisciplinary teams
what are some key characteristics of systemic reviews?
- clearly stated title/objective
- explicit criteria for inclusion/exclusion
- presentation of characteristics of each study and analysis of method quality
what are some steps in the process of conducting a systemic review?
- define review questions
- develop protocol
- identify relevant studies
- assess eligibility
- extract data
- critically appraise
- synthesise
- disseminate
what will a well-formulated review question encompass?
- search strategy
- inclusion/exclusion
- data extraction
- choice of synthesis method
- presentation/dissemination of findings
what are some challenges to lit search?
- database bias
- publication bias
- english language bias
- citation bias
when does citation bias occur?
when studies with significant or positive results are referenced on other republications, compared with studies with inconclusive or neg findings
what are the three broad categories of critical appraisal of qual studies?
- rigour: thorough and appropriate approach been applied to key methods
- credibility
- relevance
what are some methods for quant data synthesis?
meta-analysis and narrative synthesis
what are some methods for qual data synthesis?
methods depending on review purposes eg. meta-ethnography, thematic synthesis
what is a meta-analysis?
set of stat techniques for combining results from independent but similar to derive an overall estimate of treatments effect
what are some potential issues of meta-analysis?
publication bias and varying quality of studies
what are the two statistical models of analysing data used?
fixed and random effects
what is the fixed effect as a stat model to analyse data?
assumes that true effect of treatment is the same for every study
what is the random effect as a stat model to analyse data?
assumes that true effect estimate for each study vary
what are integrative reviews?
provide review of available lit according too clear methodological approach
what are some potential sources of bias for integrative reviews?
- incomplete lit search strategy
- errors in data extraction
- inappropriate data analysis
what are some lit search issues in integrative reviews?
- obtaining needed lit often challenging
- publication bias
- identify max no.of primary sources through multiple strategies
what are some data evaluation issues in integrative reviews?
- explicit assessment of study quality
- how to define quality of non-empirical studies
what are some major types of qualitative research?
- participant observation
- interview
- record review
- focus groups
- case studies
what are some stages of qual evidence synthesis?
- formulate questions
- conduct systemic lit search
- screen/select appropriate research articles
- analysing findings
- maintaining quality control
- present findings
what are some underlying assumptions of qual research?
- expect more from meta-synthesis than single
- findings that make-up a MS database have already bee saturated
- redundant saturation of exisiting findings
- priority is to synthesise more info
what is the relationship of stats to clinical practice?
- hypothesis generated
- research plan devised of study effects of treatment method
- quantifiable data collected
- results interpreted
- conclusions drawn/shared
what is the null hypothesis?
statement of no difference between/among participants
what is the alternate hypothesis?
statement that there is a definite difference among participants
what is statistical analysis concerned with?
organisation/interpretation of data according to well defined, systemic procedures
what are the 4 levels of stat analysis?
- descriptive
- inferential
- associational
- difference
what is descriptive stats as a level of analysis?
data reduction, reduces large sets into interpretable forms
what is inferential stats as a level of analysis?
inference to known population from findings in sample
what is associational stats as a level of analysis?
causality refers to set of procedures to identify relationships between multiple variables
what is the difference as a level of analysis?
determining the difference between two groups or sets of data.
what are some types of descriptive stats?
- measures of central tendency
- measures of variability
- bar graphs
- histograms
- bivariate descriptive stats
- scatter diagrams
what are some measures of central tendency?
mode, median and mean
what are measures of variability?
- range
- interquartile range
- sum of squares
- variance
- standard deviation
what are bivariate descriptive stat types?
contingency tables and correlational analysis
what does the frequency distribution reflect?
simple count of how frequently each value of a variable occurs in a set of measured objects
what are some types of distributions of frequencies?
- normal
- non symmetrical
- kurtosis
what is a normal distribution type of frequency?
both halves are identical
what is a non-symmetrical type of frequency?
either positive or negatively skewed (positive curve high on left)
what is kurtosis as a type of frequency?
characterised by shape, flatness= platykurtic, peakness= leptokurtic
what do the measures of central tendency provide info about?
most typical or representative scores in group
what is mode?
value that occurs most frequently
what is median?
midpoint
what is mean?
average
what is the interquartile range?
refers to range of middle 50% of participants
how is the sum of squares calculated?
by squaring differences each score and the mean, larger the SS the greater the variance
what is the standard deviation?
dispersion method, indicates average deviation of scores around mean
what is bivariate stats?
looking for relationships among two or more variables
what are contingency tables?
two dimensional frequency distribution that is used with categorical (nominal) data
what does correlational analysis determine?
relationship among variables
what are the three types of directional relationships that exist in correlational analysis?
positive, negative and zero correlation
what is the level of significance?
statement of the expected degree of accuracy of findings based on the sample size and on convention in literature
what does the level of significance help to ensure?
ensure stats relationship is not caused by chance
what does inferential stats provide action processes for?
drawing conclusions abut pop based on data from sample obtained
what is stat inference based on?
probability theory
what does a selection of a significance level indicate for a researcher?
indicates how confident they are that findings aren’t attributed to chance
what are type one errors in stat inference?
errors by failing to accept null hypothesis when true
what are type two errors in stat inference?
occurs in null hypothesis is mistakenly accepted when it should be rejected
what are the three assumptions of mathematical formulas that test the hypothesis of parametric stats?
- sample derived from pop with normal distribution
- variance is homogeneous
- data is measured at interval level
what is homogeneous as an assumption of parametric stats?
is displaying scores in one group having approx. the same degree of variance as the scores in another group
what are the three frequently used parametric stats?
- t-test
- one-way analysis of variance
- multiple comparison
what is t-test as a parametric stat?
compares two sample means on one variable
what are the two types of t-tests as a parametric stat?
one type for independence or uncorrelated data other for dependent or correlated
what is the one-way analysis of variance as a parametric stat?
designed to compare sample group mean to determine whether the difference can be inferred in pop, can manage two or more groups
what does the Pearson correlation coefficient measure?
strength and direction of a linear relationship between X and Y variables
what do nonparametric stats test?
tests hypotheses when they violate one or more of the assumptions for parametric procedures
what type of data is nonparametric stats used?
ordinal
what are some examples of nonparametric stats?
- chi-squared test
- mann-whitney u test
- wilcoxin signed rank test
- kruskal-wallis test
- kappa
what is chi-squared test as a parametric stat used for?
when data is nominal and when computing mean is not possible
what is mann-whitney u test as a parametric stat used for?
tests differences between two independent groups, but is used with data that are ordinal
what is wilcoxin signed rank test as a parametric stat used for?
alternative to t-test for correlated samples, at least level of ordinal
what is kruskal-wallis test as a parametric stat used for?
alternative to one-way ANOVA when comparing three pr more independent groups, ordinal
what is kappa test as a parametric stat used for?
determines degree of agreement between two or more judges independently ranking variable, categories are nominal
what are some potential sources of error affecting stat significance?
- hawthorne effect
- placebo
- honeymoon effect
- research bias
- test administrator bias
- sampling error
- systemic variance
- error variance
what is the hawthorne effect as a source of error?
attention given to subjects increases positive outcome and camouflage true effects of independent variable
what is the placebo effect as a source of error?
suggestion that subjects are receiving treatment may produce positive expectations that it will be effective
what is the honeymoon effect as a source of error?
short-term effect of new treatment that subjects hope will impact
what is the research bias as a source of error?
researcher carrying out treatment affects results through interest in it to be effective
what is the test administrator bias as a source of error?
individual administering test is aware of which subjects are in treatment group and control group
what are the sampling errors as a source of error?
results are based on sample that is not representative and will cause results to be bias or skewed
what is the systemic variance as a source of error?
fails to control extraneous variables that could impact results
what is the error variance as a source of error?
overlooks effects of anxiety, motivation, inattention and unexpected problems
what are the two aspects of validity?
internal and external
what is internal validity in qual research?
procedures used measure that they were supposed to
what is external validity in qual research?
results generalised beyond immediate study
what do qual studies that address credibility include?
- prolonged engagement and observation
- data collection triangulation
- search for neg case
- member checking
- use participant words
what is transferability?
how researcher demonstrates that findings are applicable in other contexts
what is confirmability?
ensures interpretations and findings were clearly derived from data, how conclusions were reached
what is done to establish credibility, transferability, and dependability in qual research?
- ensures findings are based on responses
- eliminate potential bias
- use audit trail/memos to highlight steps of analysis
what is dependability?
extent that study could be repeated by researchers and findings would be consistent
how can researchers ensure findings are dependable?
ensure process is logical, traceable and clearly documented
what is an audit trail?
provides readers with evidence for decisions requiring rationale
how can you maintain an audit trail?
keep records of raw data and keep reflexive journal
what is reflexivity?
challenges researcher to explicitly examine how research agenda/assumptions, subject location, beliefs and emotions enter
how can memo/reflexive journals be used in qual research?
used to record and document daily logistics, decisions and to record researchers reflection of values, interests and insight about self
how can memo and reflexive journals be done?
- field notes after interview
- how attentive is participant?
- summarise answers
- what role i play?
what is the purpose of quantitative sampling?
identify participants from whom to seek info
what are some issues related to quantitative sampling?
- nature
- size
- method of selecting
what are the three steps of quantitative sampling?
- identify pop
- define sample size
- select sample; define inclusion/exclusion
what is generalisation in quantitative sampling?
extent to which results can be extended from the sample to pop
what are the four techniques of selecting random samples?
- random
- stratified random
- cluster
- systemic
what are some advantages of random sampling?
- easy
- high probability of achieving rep sample
- meets assumptions of many stat procedures
what are some disadvantages of random sampling?
- identification of all members of pop can be difficult
- contacting members can be difficult
what are some examples of stratified random samples?
gender, ethnicity, education level, religion
what are proportional stratified random samples?
same proportion of subgroups in the samples in pop
what are non-proportional stratified random samples?
different, often equal, proportions of subgroups
what are the advantages of stratified random samples?
- more precise sample
- used for both proportional and non-proportional
- ensures rep of subgroups in sample
what are the disadvantages of stratified random samples?
identifying:
- all members of pop can be hard
- members of subgroup can be hard
what is cluster sampling?
selecting participants using groups that have similar characteristics and in which participants ca be found
what are clusters in cluster sampling?
locations within which an intact group of members of pop can be found
what are some examples of clusters?
- neighbourhoods
- school systems
- community agencies
what does multistage sampling involve?
use of two or more sets of clusters
what are some advantages of cluster sampling?
- useful when pop is large or spread over geographic region
- convenient/expedient
- don’t need names of everyone
what are some disadvantages of cluster sampling?
- rep is likely to be an issue
- assumptions of stat procedures can be violated limiting analyses
what is systemic sampling?
selecting every 5th or 10th (or some other) participant from list of members of pop
what is an advantage of systemic sampling?
easily done
what is a disadvantage of systemic sampling?
- susceptible to systemic exclusion of some subgroups
- some members of pop don’t have an equal chance of being included
when in selecting non-random samples useful?
when pop cannot be described or logistics of completion of study do not allow for random sampling
what are the three techniques of selecting non-random samples?
- convenience
- purposive
- quota
what is convenience sampling?
selection based on availability of participants
what is purposive sampling?
selection based on researcher experience of individuals being sampled
why is purposive sampling usually selected?
- knowledge that they have a diagnosis
- common experience
- being in a specific setting such as clinics etc.
what are some concerns of convenience sampling?
related to representation and generalisability are present
what are some concerns of quota sampling?
accessibility, representation, and generalisability are present
what are some concerns with purposive sampling?
concerns related to representation and generalisability are present
what is selection based on in quota sampling?
selection based on exact characteristics and quotas of participants in the sample when impossible to list all members
what are some goals of qualitative research?
- meaning
- context
- process
- reasoning
what does qualitative data provide in-depth info on?
- thoughts/views/interpretation
- priorities
- processes
- intended effects of actions
- feelings/experience
what are some challenges of collecting qualitative data?
challenging and time consuming also no step process which can create frustration
what is thematic coding?
theory emerges out of data rather than being developed in advance
what is memo writing?
make field notes throughout the entire data collection allows evaluation and modification
what are some considerations of analysis?
- chronology
- key events
- various settings
- process
- issues
what are the three key processes of grounded theory?
- initial/open coding
- axial coding
- selective coding
what is initial/open coding in grounded theory?
literal line-by-line reading and interpretation of salient categories
what is axial coding in grounded theory?
groupings of the first order, or open codes into coherent categories and subcategories
what is selective coding in grounded theory?
selection and validating major categories that outline relationships and interactions between codes
what is consensus coding?
measures the reliability of coding and improves consistency and quality of analysis
what are the two processes of inductive content analysis?
open coding and creating categories
what is open coding in inductive content analysis?
literal interpretation and making notes and headings in text
when is deductive reasoning often used?
in cases where researcher wishes to retest existing data in new data