SFL 290 Final exam

We love finals week yay i am so done

Primary Source

Original report of the original study, contains all details to duplicate/replicate study, peer review

Examples of primary sources

journal articles reporting on OG research, professional conference proceedings, scholarly books

Secondary Sources

information “once removed”, summarized information about original study, some type of peer review

Examples of secondary sources

books on specific subject, review articles in journals, textbooks

General Source

provide overview of topic, not peer reviewed, broad information

Example of general sources

newpapers, trade books, magazines, textbooks

Directional Hypothesis

nature of differences in relationship is clarified

Non Directional Hypothesis

reflecting differences and relationships but nature of differences is left open

What is theory

an explanation that puts together assumptions, constructs, hypotheses, and facts in order to explain previous findings, relates findings to each other, provide direction for future exploration – structure to facts with relationship to larger design, can help interpret findings, organize data, creates predictions

General Theory

explanations for broad range of events and can be applied to variety of concerns, examines relationships and populations

Middle range theory

substantive theories, explanations regarding a particular subject area

Parts of theory

Assumptions – beliefs accepted as given or self evident, accepted as true without being tested

Concepts – important ideas in a theory (background factors, individual characteristics, relationship quality)

Hypotheses – propositions – suggests relationships or outcomes that might be expected in model is accurate

A theoretical model – tentative, identifies all important ideas in theory and how they are related, diagram

Assumption

beliefs accepted as given or self evident, accepted as true without being tested

Concepts

important ideas in a theory (background factors, individual characteristics, relationship quality)

Hypotheses

suggests relationships or outcomes that might be expected in model is accurate

Theoretical model

tentative, identifies all important ideas in theory and how they are related, diagram

Conceptual definiton

in intro, general/dictionary definition

Operationalization

taking a concept narrowing it down, the way you measure the variable, makes an abstract idea measureable

Reliability

consistency

Test-retest Reliability

when a measure is given more than once, is there consistency between the scores, statistical test – reliability coefficient (r), higher the better

Inter-item reliability

when there are multiple items in a measure, is there consistency between the individual items and total score, are items related to each other – internlal consistency, statistical test – chronbachs alpha, high the better

Inter-rater reliability

when there are multiple raters, is there consistency from rater to rater – observational research, statistical test is cohen’s kappa

Alternate parallel form

when there are multiple forms of the same measure, is there consistency between forms of the test, statistical test – correlation coefficient (r)

Increase reliability

increase number of observations, eliminate unclear, standardize tet conditions/instruction, maintain consistent scoring procedures, moderate degree of difficulty, minimize external effects

Validity

Accuracy

Face Validity

does instrument appear to measure what it claims to measure (established by experts reviewing instrument)

Content validity

Convergent validity

does you measure hang with measures of similar ideas, typically established by correlating new with existing measures

Divergent validity

does you measure differentiate between groups of people with certain characteristics

Concurrent Validity

does your instrument correlate with current performance? Correlation coefficient

Predictive validity

does your instrument predict future performance? correlation coefficient

Types of probability sampling

Simple random, stratified, cluster, systematic

Simple Random Sample

need list of entire population, randomly select

pros - ideal, low bias

cons - time, money

Stratified

Divide population into subgroups that are mutually exhaustive and exclusive then select equal number of subjects from each subgroup

Pros - close to representative, fastish, reduced sampling error

Cons - difficult analysis, expensive, not always possible to get list of entire population

Cluster

Separate into groups then randomly select clusters and sample entire cluster

Pros - efficient, don’t need list of entire population

Cons - not always representative of entire pop. if clusters are biased

Systematic

Need list of entire population, then select every Kth individual

pros - low bias

cons - not everyone has each chance of selection after first chosen, could interact with patter in population

Non probability samples

qouta, convenient, snowball, purposive

Qouta

Sample until reach needed number

Pros - ensures representation of specific groups, low confounds

Cons - not generalizable, need qouta frame to rep. pop.

Convenient

sample accessible convenient people

pros - quick and easy

cons - bias, not generalizable

Snowball

pass along to collect sample

pros - reaches hard to reach populations

cons - time, sample may be homogenous, hard to generalize

Purposive

sample selected according to predetermined criteria

pros - reaches hard to reach populations, choose participants based on aims of research

cons - sample doesn’t represent population, vulnerable to error in judgement by researchers

Nominal

categorical, mutually exclusive categories wiithout numerical properties

examples of nominal LOM

gender, marital status, race

ordinal

ranking along continuum, more or less comparisons, distance between variables is unkown

examples of ordinal LOM

rankings, economic status, educational levels

interval

numerical with equal intervals - no true zero

examples of ordinal LOM

temp., IQ test, SAT/ACT, GPA, likert-type scales

Ratio

numerical with equal intervals, absolute zero is indicative of absence

ratio examples

age, number of years in X, number of minutes

Deduction

reasoning proceeds from a general theory to particular data, used in quantitative research and in real life

Induction

reasoning proceeds from particular data to a general theory

Triangulation of sources

1.        building knowledge using variety of sources

Triangulation of methods

mixed methods research

methods of qualitative data collection

interviews, focus groups, observations, documents, ethnography, combination, case-studies

Case studies

multi-faceted understanding of complex in real context – data – interviews, observations, medical records

Trustworthiness

rigor of design, researcher credibility, believability of findings, applicability of research

Reliability in qualitative

soundness of research/consistency of methodological process

Validity in qualitative

accuracy of findings from standpoint of researcher, participants, consumers

Credibility

Analyses are believable

Transferability

transferrable to other contexts

Dependabiltiy

anaylses are consistent and repeatable

Confirmability

analyses are supported by data

Order of qualitative data

1. open coding

2. axial coding

3. selective coding

Open coding

initial pass through data locating themes and naming theme

axial coding

going in again and focusing explicitly on themes, adjust as necessary

selective coding

looking for cases that exemplify themes, finding best qoutation

Validity: Triangulation

multiple technigues to ensure accurate description and presentation of findings

Validity: Prolonged engagement

increased amount of time in field for deeper understanding

Validity: Thick, rich, in-depth description

demonstrates embeddedness and awareness of field studied by researcher

Validity: Negative case analysis

search for and try to explain cases that don’t fit to continually revise hypothesis

Validity: Audit Trail

theoretical memos that included detailed descriptions of how you went from data to conclusions

Validity: Conceptual saturation

collect data until no new categories appear

Validity: Member check

sharing data, findings, interpretations with participants

Validity: Peer debriefing

present your analyses to other researchers to ensure own biases didn’t interfere with interpretations

Validity: Explicit documentation

documentation of data collection methods, analysis, field decisions that altered any strategies or focus

Belmont Report

Statement of ethical principles upon which federal regulations for the protection of human subjects are based

Points in Belmont report

1. Respect for persons

2. Beneficence

3. Justice

Classic experiment

Classic – control group to compare treatment against, random assignment, treatment/manipulation

Internal validity

confidently attribute results of experiment to IV, no confounds or lurkers, ensure with random selection, random assignment, and control conditions

Threats to internal validity

• History – other events occurring same time as treatment

• Maturation – biological/physical changes over time

• Selection – characteristics related to study can be used to select participants

• Testing – pretest might affect performance on later measures

• Instrumentation – how instruments are scored or used might be producing results rather than IV

• Mortality – drop out

• Regression to mean – Extreme are rare and closer to the mean when retested due to probability

External validity

confidently generalize results to other people and settings

Threats to external validity

• Multiple treatment interference – if subjects are receiving any other treatment known or unknown to researcher

• Reactive arrangements – Hawthorne effect

• Experimenter effects – behavior of researcher influencing results

• Pretest sensitization – pretesting changing nature of treatment, effectiveness change

Quasi-experimental design

control/comparison group, non-random due to ethics, treatment/manipulation

Cross-sectional

assess group differences at a single point in time (age differences rather than age changes)

Longitudinal

assess change in behavior for one group at multiple points in time

Longitudinal: Trend

changes in population over time

Longitudinal: Cohort

People with common characteristic over time, i.e. birth year

Longitudinal: Panel

group of people over time

Correlational

find relationships, not cause and effect

Differences between experimental and correlational design

• Amount of control the researcher has

• Claims you can make

• Associations vs causality

Similarities between experimental and correlational designs

• relationships between variables

• used to test theory and gather data

• IVs and DVs

Descriptive statistics

describe patterns among variables (measures of center, spread, correlation)

Mean

average

median

middle with chronological

mode

most frequent

inferential statistics

use of sample statistics to make inferences about populations

types of inferential statistics

Test-statistic, ANOVA, correlation, Regression models

T-test

compare means of two groups on same variable difference in means variation is due to chance or systemic (Ho – no difference / groups have equal means), higher t is greater differences

F-test / ANOVA

compare means across two or more groups, IV categorical, DV continuous, larger the greater the difference, post hoc tests required if more than two groups, if significant, atleast one group is different from the others

R

examine clarity and direction of relationship between two variables, IV and DV are continuous,  bigger abs. value is clearer/stronger relationship, (+) is variables move in the same direction, (-) is variables move in different directions