Research Methods PDF

Preview

Summary

Research Methods Social Work...

Description

August 30th, 2018 ● Parallels Between Social Work Research & Social Work Intervention ● Agreement reality - knowing things by belief and agreement ○ Tradition ○ Authority ● Experiential reality - knowing things by direct experience ● Errors in causal inquiry ○ Inaccurate observation - not knowing color of car next to you ○ Overgeneralization ○ Selective observation - only seeing part of it ○ Made up information ■ Ex-post facto hypothesizing (after the fact); not reason or theory based on it ○ Illogical reasoning ○ Ego involvement - one is putting themselves in the best light ○ Premature closure of inquiry - “snap decision”, cut off whole picture ○ Mystification - like believing in karma ● Theories describe logical relationships that appear to exist in the world ● Quantitative research methods emphasize the production of precise and generalizable results ● Qualitative research methods are more likely to tap into deeper meanings of particular human experiences and generate theoretically richer observations that are not easily reduced to numbers September 6th, 2018 ● ● ● ● ● ●

● ● ● ●

IRB - Institutional review board ○ All universities and hospitals that do research are required to have one Consent is a legally binding word - being treated as an adult Voluntary/informed consent Assent - a willingness to do something No harm to the patient Anonymity and confidentiality ○ Anonymous - you can not find out where it came from ○ Confidential - Knows who wrote it but do not have to tell because it is not legally necessary Deceiving subjects Analysis and reporting NASW - Code of Ethics Ethical dilemmas = weighing benefits and costs in light of ethical principles and

standards ○ Determine whether there is an ethical issue or dilemma ○ Identify ethical standards, as defined by NASW, which are being compromised ○ Weigh ethical issues in light of key SW values and principles as defined by NASW ○ Suggest modifications to research design in light of the prioritized ethical values and principles that are central to the dilemma ○ Implement modifications research design in light of prioritized ethical values and principles ○ Monitor research activities for nw ethical issues or dilemma ● Politics of social work research - concerned with practical costs and use of research (no codes to guide political conduct) ● Dual relationship - an opportunity or other relationships with clients. ○ Like a client also goes to your church ○ Try to avoid dual relationships ● Culturally competent research ○ Need to keep in mind age, gender, race, religion, sexual identity of those involved September 13th, 2018 ● Evaluation purposes ○ Formative: assessing implementation ○ Summative: assessing outcomes September 20th, 2018 ● Cross sectional - observations at one time period ● Longitudinal - observations over an extended period ○ Trench ○ Panel ○ Cohort ● Types of variables: ○ Independent ○ Dependent - affect ○ Extraneous and control ● Hypotheses - a statement that postulates or describes the relationship believed to exist between an independant and dependant variable ○ One tailed - the more experience you have the higher the salary, women are smarter than men ○ Two tailed - years of experience relates to salary, gender is related to intelligence ● Relationships between variables - will talk more about when we get to 12

○ Positive - dependent variable increases the independent variable increases or decreases as the independent variable decreases (going in the same direction) ○ Negative - two variables move in opposite directions ○ Curvilinear - the nature of the relationship changes at certain levels of the variables ● Levels of measurement: ○ Nominal (basic) - defined in terms of qualitative attributes (answers) that are categorical only ■ Are you catholic? Yes v No (qualitative difference between the answers) ■ Gender, race, religion, major ○ Ordinal - can be rank-ordered in that different attributes (answers) represent relatively more or less of the variable ■ Rating on amazon on star measurement ○ Interval - differences between levels have the same meanings ○ Ratio (highest level) - same as interval measures but with a true zero point ■ Rank order it and differences between answer categories have meaning ■ How old are you? Oldest to youngest (rank order), quantify difference between ranges (40 and 10 year old - 30 year difference), true, and absolute zero point available ● Sources of measurement error: ○ Systematic or constant ■ Acquiescent response set ■ Social desirability bias ○ Random: September 27th, 2018 ● ● ● ● ● ● ● ●

Nominal - major, sport, fave color, NO NUMBER Ratio - age, credits you’re in, height (raw #) Interval Ordinal - how satisfied are you, amazon star rating Integrity of the data - reliability and validity Reliability - consistency Validity - accuracy Sources of measurement error: ○ Systematic or constant: ■ Acquiescent response set ● Survey through 1 to 5 stars and you like the class so you go through 5 stars down the whole list. Predetermined notion so you do not read the list thoroughly ■ Social desirability bias

● Saying you studied more to make yourself look good ● Is systematic or constant because it affects everyone in the same way (maximize study time or minimize drinking)

●

● ● ● ●

○ Random: ■ Boring surveys, not knowing exam answers - make random errors ■ Filling out random answers Methods for testing reliability: ○ Inter-observer - most of the info you get is from observation ; two observers get same finding ○ Inter-rater - two people rate the thing the same ○ Test-retest - Someone asked if they were bullied as a kid once and then a week later they get the same, they should be the same bc your childhood is already over ○ Parallel-or alternate-forms - using two different measures // drunkenness - to figure out if someone is drunk you look at both slurring and stumbling both showing you they are in fact drunk ○ Internal consistency (split-half) - verbal ability on SAT split up, you could expect the student to do equal on both halves Face - does the measure appear to assess the subject matter under consideration? Content - does the measures adequately assess the major dimensions of the subject matter under consideration? Criterion related - does the measure predict future occurrences or correspond to another indicator of the concept under study? Construct - does the measure produce assessments which coincide with what is already known? Does the measure assess the desired theoretical concept?

October 4th, 2018 ● Secondary (existing) data ○ Advantages: ■ Cost ■ Speed ■ Prior testing ○ Disadvantages: ■ Not in depth ■ May limit what you see ● Observation ○ Structured ■ Not much room for interpretation ○ Unstructured ■ Ethical considerations ■ Participant observation ● Reactive effect

● Types of Data Collection Instruments ○ Interview schedule for structured/semi-structured ○ Questionnaire ■ Self-administered ● Interviews ○ Structured ○ Semi structured ○ Unstructured ○ Group November 1st, 2018 ● ● ● ●

SW doesn’t use pretest posttest control group designs bc we do not randomize Solomon four group design - rarely used Simple interrupted time series design Matching - comparison group similar to who you are studying ○ Study shu students so you won’t compare us to harvard ● External validity - ability to generalize beyond the study conditions November 15th, 2018 *For final assignment*- Designing the evaluation of practice (single subject designs) ● Researching one on subject ○ I.E. charting on behavior of one person ● Good: ○ Good indicator on is the person changing (like due to therapy) ● Bad: ○ External validity (generalizability) - what works for one does not work for all ● Self monitoring: ○ CBT - have “homework” ● Baseline phase - enough repeated measures are taken to make it unlikely that other environmental conditions are impacting on target behavior ○ Stability should be found prior to treatment introduction ● “A” is a baseline measurement ● Intervention phase: ○ Repeated measures are taken to see if the target behavior or dependent variable increases, decreases or stays the same ● Withdrawal phase: ○ Return to baseline upon withdrawal December 12th, 2018

● Possible code cleaning ● Deficiency cleaning -...