Final Review Content PDF

Midterm Review: Cri.cal thinking & Logical Reasoning - Deduc.ve reasoning o Downward (top-down) o The process of developing specific predic.ons (hypotheses) from general principles - Induc.ve reasoning o Upward (boGom-up) o The process of reasoning from specific observa.ons to more broad general rules...

Midterm Review: Cri.cal thinking & Logical Reasoning - Deduc.ve reasoning o Downward (top-down) o The process of developing specific predic.ons (hypotheses) from general principles - Induc.ve reasoning o Upward (boGom-up) o The process of reasoning from specific observa.ons to more broad general rules. o Produces theories Research is an itera.ve process: - Process involves a lot of back-and-forth movement between decision - Decisions and elements of research are interdependent and must fit together - Each decision has implica.on for other elements of the research - Decisions must complement one another if study want to generate high quality informa.on Steps of a study: 1. Significance/compliance of the study – iden.fy the problem/ develop a topic o Do a literature review to become familiar with the current state of knowledge and to iden.fy any gaps 2. Design the study o Iden.fy a methodology, popula.on/ sample, method of data collec.on, ethical principles, and demonstrate rigour/trustworthiness 3. Conduct the study o Data collec.on 4. Analyze the data o Sta.s.cal analysis (qualita.ve) o Thema.c analysis (quan.ta.ve) 5. Interpret the findings o Linking findings to research ques.on o Focuses on meaning of the results o Recognize strengths and limita.ons of study 6. Disseminate the finds – knowledge transfer o Share data with peers and interested groups/ conferences/presenta.ons Approaches to Research: 1. Quan.ta.ve (Numbers) o based on the perspec.ve that reality is objec.ve and can be measured o Uses data and sta.s.cal analysis o Uses deduc.ve reasoning (top down) – theory – hypothesis – observa.on – analysis - confirma.on o o o o Realism ontology Large sample size Focuses on parts rather than wholes Par.cipants may be in controlled, par.ally controlled, or natural se\ngs 2. Qualita.ve (Experiences) o Based on the perspec.ve that reality is subjec.ve and best represented by a collec.on of narra.ve data o Uses induc.ve reasoning (down up) – observa.on – tenta.ve conclusions – conceptual frameworks – theory o Rela.vism ontology Paradigms - A world view or belief system - In research, the paradigm determines how we ask research ques.ons and how we carry out the research - A set of assump.on, concepts, values, and prac.ces that cons.tutes a way of viewing reality There are several components of a research paradigm that are important: - Ontology – asks the ques.on, “what is reality?” o Beliefs about society/truths o Realism § A pre-social reality exists that we can access through research § Quan.ta.ve research usually § Focuses on large sample sizes to generalize it to a larger popula.on § Believes only in truth and that cannot be changed o Rela.vism § “reality” is dependent on the ways we become to know it § Qualita.ve research § Shaped by context around it § Can have more than one truth § Same sample sizes - Epistemology – asks the ques.on, “what does it mean to know?”, that is “what is the nature of knowledge? o two ways of research: § EMIC – within research study, rela.vism assump.on/ontology § ETIC – objec.ve research, realism assump.on/ontology - Axiology – asks the ques.on “what Is valuable to know?” - Methodology – asks the ques.ons “How do we find out?”, “what strategies are used?” Descrip.ons of paradigms/worldviews: - Posi4vism o this was the dominant paradigm in science and nursing for many years. o based on the assump.on that there is one, single reality (ontology) and that to know that single reality, the study of a phenomenon must be undertaken with objec.vity and detachment (epistemology). o tests hypotheses or theories using rigorous quan.ta.ve research methods with large sample sizes. o The research is detached from the research par.cipant to avoid bias and ocen is blinded to the specific condi.ons of the research study. o The objec.ve of posi.vism is to produce data that is predic.ve, generalizable to a popula.on or situa.on. o This approach uses deduc.ve reasoning. - Post posi4vism o Over .me even posi.vists recognized that human behavior is complex and that it is not always possible to be completely unbiased and objec.ve. o based on the belief that there is a single reality, but we cannot know that for sure, and a modified objec.vist epistemology, recognizing that complete objec.vity is not possible. o Think of Posi.vism as belonging in the laboratory where people work with petri dishes and post posi.vism as applying to working with people. - Construc4vism (Interpre4vist or Naturalis4c) o based on a rela.vist ontology in which reality is viewed as subjec.ve and differs from person to person, and that people construct their own reali.es. o Reality occurs within contexts so many construc.ons are possible. o uses induc.ve reasoning and can develop theories from specific observa.ons. o Qualita.ve research methods with small samples are used to collect rich, indepth data that describes the individual experiences of the study par.cipants. o The researcher using this approach recognizes that they cannot be separated from the research process or from the people involved. o The research involves interac.on between the researcher and par.cipant and uses qualita.ve approaches. o Usually the researcher acknowledges his/her biases and uses self-reflec.on. - Transforma4ve (Advocacy/Par4cipatory) o focused on raising awareness and promo.ng social change. o purpose is to empower groups of people who are marginalized, or to inves.gate inequali.es or social injus.ces. o approach it is believed that knowledge is socially and historically constructed, so the ontological posi.on is historical realism, and the epistemological posi.on is social construc.vism. o Ocen a cri.cal research approach is used guided by different ideological posi.ons such as feminist research, query theory, par.cipator ac.on research - Pragma4sm o view the tradi.onal (posi.vist/postposi.vist/construc.onism) paradigms as prescrip.ve in their approach. o focuses on problems rather than philosophical posi.oning, so aGempts to avoid a single ontological and epistemological approach. o Both quan.ta.ve and qualita.ve methods are used in a pragma.c approach. In other words, researchers “do what works” to achieve answers to research ques.ons Qualita4ve Research Emerging design Complex picture Research bias present Context/se\ng important Open ended data collected Induc.ve data analysis Flexible wri.ng structure Quan4ta4ve Research Fixed design Narrowed picture Researcher bias absent Contrived se\ng Closed ended data collec.on Deduc.ve data analysis Highly structured wri.ng Cri.cal Reading Process of Research Ar.cle: 1. Preliminary – skimming the ar.cle to become familiar 2. Comprehensive – understand the purpose/intent of research 3. Analysis – understand the parts/components of the study 4. Synthesis – understanding the whole ar.cle and each step of the research product Terminology: Research problem – a puzzling, perplexing, or troubling condi.on/situa.on that a researcher wants to address through inquiry Research statement – a statement ar.cula.ng the research problem and indica.ng the need for a study - Need to meet 3 criteria o Researchability o Feasibility o Significance - Ex) “uncertainty exists as to whether gastrointes.nal symptoms may explain the poorer treatment outcome of a women with celiac disease” Statement of purpose – the summary of the overall study goal - 2-3 sentences Quan.ta.ve: o Iden.fies key variables o Iden.fies possible rela.onships among variables o Indicates the popula.on of interest o Suggests, through use of verbs, the nature of the inquiry Qualita.ve: o Iden.fies central phenomenon o Indicates the research tradi.on o Indicates the group, community, se\ng of interest o Suggest, through use of verbs, the nature of the inquiry Research aims/objec4ves – the specific accomplishments to be achieved by conduc.ng the study Research ques4on – the specific query the researcher wants to answer in addressing ghe researcher problem - Has an ac.ve stem – use what or why - Includes the topic of interest o Ques.on, not statements o Simplicity – precise and specific o Ac.on – oriented o Provides direc.on for ac.on Components of Ques.on Types 1. Exploratory/descrip.ve – one variable/phenomenon and one popula.on topic o Ex) what are the characteris.cs of pa.ents with hip fractures? 2. Correla.onal/compara.ve – a minimum of two variables o What is the rela.onship between dietary intake and birth weight? 3. Quasi-experimental/experimental – must be two variables that specify a cause and effect o Why does preopera.ve teaching decrease postopera.ve anxiety? Hypothesis – the predic.ons about the rela.onship among variables Testability: - Correla.onal/experimental research ques.ons must imply that the problem is measurable/testable Research ques4ons should be: F – feasible - Research subjects must be available - Time and money needs to be available - Researcher needs to be interested in the study and is capable of carrying out the study I – interes.ng - Is the ques.on interes.ng - Funders are more likely to provide more money if ques.on is interes.ng N – novel - Will the study provide new evidence? - Will it fill any gaps? E – ethical - Will the study meet the requirements of the research ethical board (REB)? R – relevant - Will the study improve clinical prac.ce? - Will it improve pa.ent outcomes? - Can the results be implemented in different clinical se\ngs? Formula.ng Research Ques.ons: 1. PICOT-D Format o P – popula.on o I – interven.on o C – comparison o O – outcomes o T – .ming o D – data 2. PICo Format o P – popula.on o I – interest o C – context Defini4ons of Concepts for Quan4ta4ve research Conceptual – theore.cal defini.on - Ex) .me during which a person is a registered pa.ent at a hospital Opera.onal – how the concept will be measured - Ex) sum of days as a registered pa.ent, beginning with admission day and concluding with dismissal day Quan.ta.ve variables - Independent variable – the cause or influencing variables o Intake of caffeine - Dependant variable – response/ outcome o Wakefulness - Extraneous/ confounding variables – exisit in all studies and can limit a study’s usefulness o Can affect the measurement of the study variables and rela.onship among them o Classified as recognized/unrecognized and controlled/uncontrolled Quan4ta4ve Research Design Cri.cal criteria for quan.ta.ve designs: - Objec.vity - Accuracy - Control Feasibility Validity – to what extent are the findings believable/true 1. Internal Threats to internal validity – unreliable findings about the dependant variable o Features of the research situa.on that compete with the independent variable as an explana.on for what is observed about the dependant variable § History – external events that occur at the same .me as the independent variable that can affect the outcome § Matura.on – processes that occur within the par.cipants of a study as a result of .me § Tes.ng – tes.ng effects occur in designs that have more than one stage. Taking a test more than once influences the behaviours and scores § Instrumenta.on – changes in data collec.on methods § Mortality – par.cipants drop out or die § Selec.on bias – the way people are chosen to par.cipate in a study 2. External Threats to external validity – deals with possible problems of generalizability of the inves.ga.on’s findings to addi.onal popula.ons and to other environmental condi.on o Selec.on effect o Reac.on effect o Measurement effect Categories of quan.ta.ve designs 1. Experimental Features: o Interven.on – experimenter controls the IV and observes the effects on the DV o Control – researcher introduces controls into study o Randomiza.on – random assignment to groups Designs: o Basic acer only design § Random assignment to two groups o Before acer design § Random assignment to two groups o Solomon four group design § AGempts to control for the effects of pre-tes.ng by adding groups that are not pre-tested § A combina.on of the acer only and pretest/post test design o Crossover design § Researcher exposes the same study par.cipants to more than one interven.on o Factorial design 2. Quasi-experimental o Similar to experimental but lack randomiza.on o Non-equivalent group design § BXA § BCA o Time series § B1B2B3 x A1A2A3 3. Non-experimental o Case study o Surveys o Descrip.ve comparison o Descrip.ve correla.on o Correla.on designs § Case-control (retrospec.ve) § Cohort (prospec.ve) When to use Qualita.ve Approaches 1. Explora.on o Purpose is to understand, describe, explore contexts o If there is liGle to no research done on a topic, qualita.ve research helps to establish a base of knowledge 2. Complexity o Some.mes quan.ta.ve research isn’t able to unravel the complexity of a phenomenon 3. Varia.on between par.cipants o A great varia.on between responses of par.cipants in a quan.ta.ve study, the researchers might conduct a qualita.ve interview to tease out reasons for varia.ons 4. Inadequacy of conven.onal theories o When it is not possible to prove a theory by using quan.ta.ve methods, researchers will conduct a qualita.ve study to try and generate a new theory Qualita4ve Designs 1. Ground theory o Based on observa.on of the world as it is loved by a selected group of people o A theory about basic social processes o Method – aim of genera.ng a theory that explains a context or phenomenon o Founda.on – sociology, social psychology, interac.on theory o Data collec.on – interviews 20-30 people who par.cipate in a process about a central phenomenon o Data Analysis – § Step 1 (Open-coding) – take raw data or transcripts and segment them into categroies of informa.on. Develop categories to revel central phenomenon § Step 2 (Axial coding) – return to data and examine the central phenomenon in more detail by asking a series of ques.ons such as “what caused this phenomenon, what strategies were used in response to it?” 2. Ethnography o Designed to produce cultural theory o Scien.fically describes cultural groups o Goal is to understand different views of the world o Founda.on – anthropology Two perspec.ves: a. Emic (insider POV) - very ocen the researcher would be a member of the community being researched b. E.c (outsider POV) - the researcher is a person who is from a different community Combina.on of the two perspec.ves = richest interpreta.on 3. Phenomenology o Aims to describe experiences as it is lived o A descrip.on of an experience as it is lived o Method – understanding the essence of a lived experience/phenomenon o Founda.on – philosophy o Data collec.on – long interviews with up to 10 people who have experienced a phenomenon o Purpose – develop a richer understanding of human phenomenon Two Approaches: c. Descrip.ve – focus is on descrip.on of the experience to get its true meaning. Uses bracke.ng (researcher puts aside their own beliefs and biases about the phenomenon being studied) d. Interpre.ve – does not use bracke.ng. The researcher included their own experiences and perspec.ves into the study 4. Historical o Systemic complica.on of data and the cri.cal presenta.on, evalua.on, and interpreta.on of facts regarding people, events, and occurrences of the past o Founda.on – philosophy, art & science o Oral history – collec.ng data on past events by interviewing knowledgeable person about the event 5. Narra.ve inquiry o Stories shape and construct experiences o Provide a way of understanding how people interpret certain situa.ons and create reality that they act upon 6. Case study o In-depth contextual analysis of an en.ty or a small number of en..es – can be individual, family, ins.tu.onal or other social groups Sampling – Quan4ta4ve Research - Popula.on o Can be people, animals, objects, events o A well defined set that has certain proper.es - Target popula.on o The en.re set of individuals or elements who meet the sampling criteria - Accessible popula.on o Is the por.on of the target popula.on to which the researcher has access - Sample o Set of elements that make up the popula.on Sampling Plans 1. Probability a. Simple random sample § Uses a sampling frame – a list of all popula.on elements § Involves random selec.on of elements from the sampling frame b. Stra.fied sample § Popula.on is divided into two or more strata, than select elements randomly from each subgroup § Dis.nguish popula.on members from one another – gender, social class, educa.on level c. Cluster sample § Successive randomly sampling of units from larger to smaller units – ex) provinces, postal codes, households § Used when the popula.on is large and widely dispersed d. Systema.c sample § Selec.on of every nth case from the available popula.on using a random star.ng point § Advantages – avoids researcher bias, can es.mate sampling errors § Disadvantages .me consuming, inconvenient Random selec.on (RS) – refers to how to researcher selected the sample from the popula.on Random Assignment (RA or R) – refers to how the subjects were assigned to either the treatment or control group in a study 2. Non-probability a. Convenience § Use of the most conveniently available people b. c. d. e. § Most widely used approach by quan.ta.ve researchers § Most vulnerable to sampling biases Quota § Convenience sampling within specified strata of the popula.on § Enhances representa.veness of sample Matching § Used to obtain equivalent groups § Subjects are matched on some characteris.c that could affect the dependant variable Consecu.ve § Involves taking all of the people from an accessible popula.on who meet the eligibility criteria over a specific .me interval or for a specific sample size § Risk of bias low Purposive § Sample members are hand picked by researcher to achieve certain goals § Used more ocen by qualita.ve than quan.ta.ve researchers Factors influencing sample size - Type of design used - Type of sampling procedure used - Heterogeneity of the aGributes under inves.ga.on - Rela.ve frequency of occurrence of the phenomenon of interest in the popula.on - Projected cost of using a par.cular sampling strategy Sampling Problems 1. Sampling errors - Differences between sample values and popula.on values - The fluctua.on of the value of a sta.s.c from one sample to another drawn from the same popula.on 2. Sampling bias - Over or under representa.on of segments of the popula.on on key variables when he sample is not representa.ve 3. Sample size - The number of study par.cipants in the final sample - Sample size can be determined through power analysis - Sample size is a key determinant of sample quality in quan.ta.ve research Quan.ta.ve Data Analysis Level of measurement 1. Categorical a. Nominal – very simple analysis – what colour is your hair? Male/female? What is your na.onality? b. Ordinal – used for measurement of sa.sfac.on. How likely are you to do “blank”? 2. Con.nuous o Parametric tests a. Interval – you are able to have nega.ve number and the value of zero has meaning ex) temperatures. Can calculate the mean. b. Ra.o – similar to interval but do not have nega.ve numbers. Ex) height or weight. No absolute zero, as zero would mean nothing is there. Can calculate the mean. Types of Analysis 1. Descrip.ve Analysis - Analyses one variable at a .me (univariate analysis) - To describe the characteris.cs of those involved in a study using the following three elements: o Frequency & disturbance o Central tendency o Dispersion/measures of variability 2. Inferen.al Analysis - Used to aGempt to establish If there is a rela.onship between variables. - Can be applied to experimental, quais-experiemental, and non-experimental studies that focuses on two or more variables - 2 main classifica.ons o Parametric § More sensi.ve and produce more powerful results than non-parametric. § Needs to be interval or ra.o data (con.nuous measurement) § Assumes normal distribu.on of variables § Par.cipants should be randomly assigned § Homogeneity of variance (popula.on variance of 2 or more samples are considered equal). § Tests of differences: t-tests (independent, correlated, ANOVA, MANCOVA) • Assump.ons: o Independent variable is nominal level indica.ng group status o Dependant variable is interval or ra.o o Assumes random sampling o DV is normally distributed within each popula.on (sample) o Assumes homogeneity of variance in popula.on (sample size) § Tests of associa.on: Pearson r o Non-parametric § No rigorous assump.ons about distribu.on of the variables § Used with small samples § Uses nominal or ordinal data (categorical) § § § When distribu.on is severely skewed Tests of difference: Chi-square Test of associa.on: Phi coefficient, spearman Significance level/probability P-value: - Expressed on a scale from 0-1 - 0 signifying there is no chance of its occurrence - 1 signifying it is certain to happen - Probability is defined as the likelihood that the results were not obtained by chance alone if the null hypothesis is true - The p value is calculated based on the assump.on that the null hypothesis is true and tells researchers how rarely they would observe a difference as large (or larger) than the one they did if the null hypothesis were true Test of Differences - Aim: test of differences between groups o More than one group o Data needs to be nominal, ordinal, interval (ra.o) Independent T Test - Rehab trial o 2 independent groups o Distance walked is con.nuous. o Most appropriate summary measure = mean o Best compara.ve summary measure = difference between the mean distance walked post rehab between the two groups Independent Group T-Tests - Independent (unpairs) group t-test – t o Subjects in 2 groups are not the same people and not connected in any systema.c way. o Ex: caffeinated coffee and intraocular pressure (n=100) § Group 1 = regular coffee n=50 § Group 2 = decaffeinated coffee n=50 § IV = group = nominal level § DV = intraocular pressure – ra.o Dependent Group T-test - Dependant group t-test = paired, correlated group t-tests) o Same group of subjects is measured on more than one occasion, sample fluctua.on is lower § Ex) diabe.cs on weight loss program Test of Differences: 1. Analysis of Variance (ANOVA) – parametric o When the mean of three of more groups are compared o Assumes independence of groups o Dependent variables should be con.nuous (Ra.o or interval) and normally distributed - Ex) Effects of body posi.on on cardiovascular response during Valsalva maneuver o 4 randomly assigned posi.on groups (IV – nominal level) § Lying on back § Sit upright at 45 degrees § Right side lying with legs bent § Sit in chair at 50 degree angle o Measure BP during maneuver (DV = ra.o level) 2. MANOVA (Mul.variate analysis of variance – parametric o Examines differences between mean scores of two or more groups on two or more dependant variables that are examined at the same .me o Repeated measures ANOVA – group means are compared at mul.ple points 3. Chi Square – Non-parametric o Used with nominal/ordinal data o Compares observed frequencies with expected frequencies o Expected frequencies are the number of cases in each case if null is true o Fischer’s Exact probability test used for small sample sizes (<6 in each cell) o E.g. Test for differences in study groups on marital status, racial makeup, educa.onal level External Validity: generalizability of research findings to other se\ngs and samples - Are the casual rela.onships between the variables applicable to other persons, places or .mes? Factors that effect these include: o Selec.on effects: selec.on of subjects § Par.cipant characteris.cs may affect the observed response to the IV in a way that the findings cannot be generalized to people with different characteris.cs § Increased threat to external validity – when cannot obtain ideal sample o Reac.ve effects: study condi.ons § Subjects may be reac.ng to some feature of the research environment rather than the IV; cannot assume that findings will be same when IV in administered under different circumstances. • Experimenter effect – E.g., Experimenter bias - Influence that the researcher might have on the result. • Novelty effect – E.g., subject experience with anxiety in the first exposure to IV then the level of anxiety decreases in subsequent exposure. • - Hawthorne effect: types of observa.on - Changes in subjects’ behavior because they know they are being observed. o Also known as reac.vity The significance of a quan.ta.ve study is highly dependent on its applica.on to more than one se\ng. Measurement Effects - pretest can increase or decrease subject’s sensi.vity to the dependent variable - affects generalizability if pretest not used in other se\ngs, popula.ons, etc. o example – based on pre-test, par.cipants modify their thinking and responses o increased threat with pre-tests Rigour In Research - important to discover the truth - we want research to have internal and external validity - the more control you have in your quan.ta.ve methodology, the more confidence you can have in the validity of the findings. Reliability - Consistency with which a measuring instrument yields a certain, consistent result when the en.ty being measured hasn’t changed - Extent that the instrument yields the same result on repeated measures - Analogous to variance (low reliability = high variance) - A reliability coefficient of r= .85 means that 85% of variability in observed scores is presumed to represent true individual differences and 15% of variability is due to random error § Weak Reliability: 0.00-0.04 § Moderate Reliability: 0.41-0.60 § Strong Reliability: 0.61-0.80 § Very Strong Reliability: 0.81-1.00 o Stability: the degree to which an instrument generates similar findings from the same (or similar) group of individuals on different occasions o Internal Consistency: the degree to which items on a ques.onnaire measure a par.cular variable. o Equivalence: of two instruments (interrater reliability) Threats to Reliability - Unstable/ unrepeatable measurements - Bias - Small sample size Tests for Reliability - Tests of Stability: Does it generate similar findings on different occasions? o Test-retest – compare results of tes.ng at two points in .me with same individuals. - - o Agreement of measuring instruments over .me (e.g. calibra.on, stable traits) o Tests of Homogeneity/ Internal Consistency: does it measure a par.cular variable with a strong internal consistency, are said to be homogenous. o Extent to which tests assess the same characteris.c or quality. o Example: a ques.onnaire contains ques.ons on anxiety; internal consistency tells us which ones focus on anxiety o Cronbach’s alpha for con.nuous data (range = 0-1). Measures propor.on of variance that is shared among items o Kuder-Richardson 20 (KR20) for dichotomous data Tests of equivalence: degree of similarity between two or more alternate forms of a measurement instrument o Similarity between two or more alternate forms of an instrument o Intra-rater reliability – assesses how one person rates same observa.on on 2 or more occasions: consistency. o Inter-rater reliability – degree to which two or more independent observers agree. o Measured by Cohen’s Kappa, Person’s r, or Spearman’s rank correla.on coefficient. Validity - Degree that the instrument measures what it is supposed to measure - Analogous to unbiasedness (valid = unbiased) - Validity is affected by systema.c error - E.g. suppose a pulse oximeter always reads 2% points below the actual blood O2 satura8on. Systema.c error will affect everyone’s readings Internal Validity 1. Ensuring internal validity relies heavily on the use of data collec.on tools or study instruments that promote accurate, precise and true collec.on of study data. 2. Consider how difficult it is to measure quality of life in an instrument – it can never be 100% accurate. 3. There are 4 main types of instrument validity: ¡ 1. Face Validity ¡ Subjec.ve evalua.on of a measurement instrument based on the way it appears ¡ Does the instrument look like it is measuring what it is intended to measure ¡ It is the weakest form of validity ¡ 2. Content Validity ¡ Does the content of the instrument adequately capture the construct? ¡ The instrument is distributed to a panel of experts under inquiry. The panel rate each item on a scale 1-4, where 1 is not all relevant; 2 is somewhat relevant; 3 relevant and 4 is very relevant. ¡ Any score less than 3 should be removed from the ques.onnaire. Based on comparing the content of the measure with the theore.cal defini.on of the construct. ¡ Subjec.ve ¡ 3. Criterion Validity ¡ indicates how well, or poorly, an instrument compares to either another instrument or another predictor. There are 2 types of criterion validity: ¡ Concurrent: is used to determine the accuracy of a data collec.on instrument by comparing it with another data collec.on instrument i.e. compare to a “gold standard” ¡ Predic4ve validity: how accurately a measurement instrument or test will predict the outcomes at a future .me. ¡ 4. Construct Validity ¡ The extent to which a test measures a theore.cal construct or trait and aGempts to validate a body of theory underlying the measurement and tes.ng of the hypothesized rela.onships ¡ Indicates how well the scale measures the construct it was designed to measure ¡ The most complex type of validity, and involves rela.ng an instrument for data collec.on to a theore.cal framework – usually involves hypothesis tes.ng ¡ Example, a researcher inven.ng a new IQ test might spend a great deal of .me aGemp.ng to "define" intelligence in order to reach an acceptable level of construct validity. o Approaches to Construct Validity § Factor Analysis: Sta.s.cal procedure to iden.fy the underlying dimensions of an instrument. • One sta.s.cal technique that can be used to determine the constructs or domains with a developing measure and therefore contributes to establishing construct validity. § Known-Groups Approach: Administer the test to 2 groups known to differ. Also called the Contrasted Groups Approach • If the instrument is sensi.ve to individual differences in the trait being measured then results should show differences § Hypothesis Tes4ng: Tes.ng of the hypothesized rela.onships. • Test can discriminate between a group of individuals known to have a par.cular trait and a group who do not have the trait. Measurement Errors ¡ Random Errors – variable ¡ an error that causes individuals’ observed scores to vary haphazardly around their true score ¡ Affects reliability ¡ 1. Environmental factors ¡ 2. Researcher factors ¡ 3. Subject factors ¡ 4. Instrumenta.on factors ¡ Systema4c Error – constant ¡ error that is not random but occurs consistently, e.g. a scale that inaccurately weighs subjects 3 pounds heavy ¡ affects validity ¡ 1. Researcher factors – observer bias, may consistently rate subjects higher/lower ¡ 2. Subject factors: response is biased – par.cipant always answers nega.vely or posi.vely ¡ 3. Instrumenta4on factors – inadequate sampling of items in domain of interest E.g. Forge\ng to clear cache before .ming a run Mixed Methods - Planned integra.on of qualita.ve and quan.ta.ve data in a single study Advantages • Complementarity: Avoids limita.ons of a singular approach • Prac.cality: Prac.cal to use whatever tool are best suited to answer the research ques.on. • Enhanced validity: If the hypothesis is supports by complimentary data, researchers can increase confidence in inferences. Reasons to Use Mixed methods • When concepts are poorly understood • Finding from one approach enhanced by the other • When one approach is inadequate • Quan.ta.ve results difficult to interpret- qualita.ve data can help with explana.ons • Mul.phase projects needed to achieve objec.ves Sequencing • Concurrent designs both strands occurring in one simultaneous phase • Sequen.al designs one strand occurring prior to and informing the second Nota.on • All capital for the dominant strand • All lowercase for the nondominant strand • An arrow is used for the sequen.al designs • “+” is used for concurrent designs • Parentheses can be used to show an embedded structure. Sampling Strategies Iden4cal – same par.cipants are in both strands Nested – some of the par.cipants from one strand are in the other strand Parallel – par.cipants are in either one strand or the other, but drawn from a similar popula.on Mul4level – par.cipants are not the same and are drawn from different popula.ons at different levels in a hierarchy Exploratory Design - a research or design approach that is focused on investigating and understanding a problem or phenomenon in an open-ended and flexible manner. This type of design is often used in the early stages of a project when there is limited knowledge about the subject, and the goal is to gather information, generate ideas, and explore potential solutions. Explanatory design - is a research or design approach that aims to provide a deeper understanding of the rela.onships between variables or to explain the causal factors underlying a par.cular phenomenon. Unlike exploratory design, which is more open-ended and focuses on genera.ng insights and ideas, explanatory design seeks to test hypotheses, establish causa.on, and provide explana.ons for observed phenomena Nested Design - one set of experimental condi.ons or factors is nested within another. In other words, the levels of one factor are not crossed with all levels of another factor but are instead treated as subgroups or subsets within the levels of another factor. This design is ocen used when there is a hierarchical or nested structure in the experimental condi.ons. Qualita.ve Research Analysis • Comprehending: Making sense of the data and learning about what is going on Require prepara.on of thorough and rich descrip.on of data • Synthesizing: Ge\ng a sense of what is typical about the phenomenon Making generalized statements about the phenomenon • Theorizing: Systema.c sor.ng of data and the development of explana.ons that fit with the data • Recontextualizing: Development of theory applicable to the se\ngs or group explored Common Features: - Data reduc.on - Reflexivity - Data collec.on and analysis occur simultaneously Steps involved in qualita.ve data analysis - - - - Management or organiza.on of data o Manual methods § As the analyst makes sense of the data, codes are sorted into categories using color codes. o Computerized Methods Becoming immersed in the data o Gaining general sense of the data o Reading and re-reading o Listening regularly to audiotapes Reducing the data o Ongoing process as data is collected o Process of selec.ng, focusing, simplifying, abstrac.ng, and transforming the data o Organized into meaningful clusters (themes or structured meaning units) o Thema.c analysis: process of recognizing and recovering the themes o Memos are kept to organize data; write personal notes to self. o Researchers use memos in the margins of transcripts and notes to record decisions they made at a stage. o Data is coded – given a tag or label according to theme/category o Coding involves applica.on descrip.ve labels § in-vivo coding – wording/terms that par.cipants use § Constructed codes – coded data from in vivo codes, codes created by researchers, academic terms o Codebook used to organize codes into lists Rela.ng themes and categories o Codes must relate to the themes o Themes must relate to each other Presen.ng the outcomes o Presenta.ons of outcomes/findings occur acer data analysis o Involves presenta.on of the themes and the codes o Quotes jus.fy the classifica.on of codes and themes o Discussion of findings (in the context of literature) o Implica.ons of findings for nursing prac.ce Ethnographic Analysis • Iden.fy paGerns in behaviors and thoughts a given culture • Compare the paGerns and analyze many paGerns simultaneously • Gain a deeper understanding of the paGerns under study • Taking cultural inventory Phenomenology Analysis - focused on describing the essen.al nature of lived experience, ocen through the iden.fica.on of essen.al themes. Ground Theory Analysis - Three levels of coding. The researcher started with: • Level I codes: In vivo codes • Used words of par.cipants in vivo codes • Level II codes: Categories • Level I codes are collapsed into Level II codes • Data moved to a more abstract level • Level III: Construct • Increase the scope of theory generated Conclusion Drawing & Verifica.on - Challenge – is for researcher to stay open to new ideas, themes, and concepts as they appear - Researcher may draw conclusions prematurely o Conclusion drawing is the descrip.on of the rela.onship between themes § Involves or related to doing abstrac.on – move data from codes and themes to concepts/constructs o Verifica.on occurs as data is collected § Rechecking the data through verifica.on by colleagues § Finding new ideas – applying the model to the new ideas. Qualita.ve Research Rigour Goals of good research: - Truth value - Applicability - Consistency - Neutrality Rigour: Extreme thoroughness and accuracy in research achieved through strict methods, processes or procedures - Similar to validity and reliability in quan.ta.ve research Rigour Evalua.on Criteria: 1. Credibility - confidence in the 'truth' of the findings Techniques for establishing Credibility: E.g. Prolonged Engagement, Persistent Observa.on, Member Checking - Need to develop a trus.ng rela.onship with research par.cipants - Need to observe and interact in various contexts over .me - Need to get a deep and complex understanding of the phenomenon under study 2. Transferability - showing that the findings have applicability in other contexts Techniques for establishing Transferability: E.g Thick Descrip.on 3. Dependability - showing that the findings are consistent and could be repeated Techniques for establishing dependability: E.g. Inquiry Audit 4. Confirmability - a degree of neutrality or the extent to which the findings of a study are shaped by the respondents and not researcher bias, mo4va4on, or interest Techniques for establishing dependability: E.g. Audit Trail, Peer Debriefing 5. Authen4city/Fi^ngness – faithfulness to the everyday reality of the par4cipant Techniques for establishing confirmability: E.g. Audit Trail, Triangula.on Triangula4on - Methodological triangula4on: Mul.ple methods of data collec.on (interviews – individual and group, observa.on, literature, archives) u Data Triangula4on: u Collect data at different .mes and se\ng u From different group of people (e.g. Interview diverse key informants such as nurses & pa.ents) u Inves4gator triangula4on: Mul.ple inves.gators u Methodological triangula4on: Use of mul.ple methods u Theory triangula4on: (mul.ple perspec.ves to interpret data set) u Interdisciplinary: The use of other disciplines Scien4fic observa4on u Consistent with study objec.ves u Systema.c plan for recording data u All observa.ons are checked and controlled u Observa.ons are related to scien.fic concepts Conduc.ng Observa.ons 1. Structured Observa4ons u Preset idea of what will be collected – may be checklists, ra.ng scales, category system u Record the frequency, dura.on and appearance of a pre-specified behaviour or event 2. Unstructured Observa4ons u More flexibility (openness) in what will be recorded u Benefits u Richer understanding of human behaviours and interac.ons u Allows for flexibility to re-conceptualize the problem u Disadvantages u Bias u Quality of observa.on depends on interpersonal and observa.onal skills