Truth in Human Development & Family Studies HD FS 503 Kevin Spacey in “Midnight in the Garden of Good and Evil”: ? Truth, like art, is in the eye of the beholder… ? You believe what you think ? And I’ll believe what I know. Lecture Notes February 4, 2002 Optimizing Criteria for Research 1 2 Assignments ? ? Sampling Frame Very interesting, thoughtful analyses Communication hints: Avoid indefinite pronouns: this, it Use past tense to describe all aspects of study except theoretical conclusions ? Use while and since ONLY for time; use although, because, for other relationships ? Distinguish between that (restrictive clause) and , which (descriptive clause) ? Remove , which clause without altering meaning ? Avoid anthropomorphizing: The authors found, NOT the study found ? Avoid quotes; paraphrase except for very memorable statements ? ? ? References: omit issue numbers 3 ? List of units from which we draw sample ? Faculty or School directory ? Telephone directory ? List of licensed child care centers in county ? Membership directory of professional organization ? Be sure that the sampling frame accurately represents the population ? Not a sampling frame of convenience!!! phone book for 1931 election!! ? E.g., Challenges in sampling Types of Sampling ? Unstable population, such that sampling ? Probability Sampling frame quickly outdated ? Population for whom no sampling frame exists ? Nonprobability Sampling ? Non-registered 4 family child care providers population ? Potentially stigmatized ? Ethnic minority ? Gays and lesbians 5 6 1 The most common probability sampling techniques Probability Sampling ? Requires random sampling of units from population ? Allows generalization of findings to population ? Selects participants or situations for study 7 Other common probability sampling techniques Nonprobability sampling Cluster sample: ? You want to study the impact of welfare reform in Iowa on mothers’ employment ? You select several ? Counties (rural and urban?) Communities (large, small?) ? Low income neighborhoods ? ? Simple Random Sampling: ? Each unit of the population has an equal chance of being selected ? Choice of any unit is independent of choice of others Proportional Stratified Random Sampling: ? For potentially moderating variables (e.g., gender, class), units chosen from each group in proportion to the group size ? Caution: Strata must be accessible in sampling frame! 8 Then do door-to-door canvassing 9 Sampling unit ? Sometimes used for convenience ? NOT a good idea if you hope to generalize findings to a population!!! ? In qualitative research, it’s up to reader to decide whether results can be transferred to another context or group ? Why nonprobability sampling? ? For rarer populations ? For samples with no appropriate sampling frame ? For VERY expensive samples 10 ? e.g., clinical samples Stratified Sampling ? The family? child? classroom? student? ? Sampling unit is smallest unit receiving treatment ? If teacher-taught curriculum, ?the classroom is the unit, not the child!! ? If computer-based curriculum ?the child is the unit ? If parent-child interaction ? the ? ? Divide members of sampling frame into strata ? Each strata groups members who share a characteristic ? This characteristic might affect results ? Examples of strata gender grade ? ethnic group ? ? dyad is the unit ANOVA allows testing at level of group (first), then individual level (nested design) ? 11 May be studied as a moderating variable 12 2 Proportional Stratified Sampling Oversampling ? Members selected from each strata ? According to distribution in population ? Multiple strata possible ? By sex within grade ? Sex and grade are potentially moderating variables (“Treatments”) ? Test for interactions among independent variables: ? for Treatment X, those who do better may be ? boys in second grade ? girls in fifth grade ? Proportional sampling will produce too small a sample to represent one strata that has few members (e.g., one ethnic group) ? Oversample strata with few members 13 14 Cluster sampling Nonprobability sampling ? Randomly select clusters (typically, ? Judgmental sampling defined by geographic proximity) ? Purposive sampling ? neighborhoods, ? Snowball (chain referral) sampling ? towns, ? Sequential sampling ? schools, ? classrooms ? Randomly select members within clusters 15 16 Purposive sampling Judgmental sampling ? NON -probability, Non-random!!!!!!! ? Special case of judgmental sampling ? Pick out certain interesting cases ? Often used in qualitative research ? Relies on your experience and Most representative cases? Cases likely to contradict previous results? ? Examples: ? High levels of education, low levels of income ? Hoffman-Reim : chose cases widely different from each other ? N.B.: All sampling is purposeful; however, random sampling is not purposive ? ? knowledge of theory and previous research findings ? 17 ? If I want to study elderly persons, I sample from the population of elderly persons, not high school students! This sampling is purposeful!!!! 18 3 Purposive sample (e.g., Hoffman-Reim) Typical randomly selected sample 6 40 Number of participants 30 20 10 0 5 4 3 2 1 Also known as chain referral sampling ? Ask one leader to identify other leaders in community ? Ask those individuals to name others ? Keep going until you return to first names ? For a difficult-to-locate population (e.g., for study on unsafe sex practices among homosexual males) ? 20 80 70 Quota sampling ? ? 60 Characteristic 19 Snowball sampling 90 10 0 Characteristic 50 30 0 10 10 20 30 40 50 60 70 80 90 100 40 0 0 20 Number of participants 50 ? Keep knocking on doors until you fill preset quotas ? single parent families ? DINK’s ? two parent families ? retired couples ? single persons More common in qualitative research May lead to biased sample in quantitative research ? Danger!!! ? convenience leads to bias 21 22 What sampling strategy have we omitted? Sampling strategy ? Convenience sample ? Choose sampling technique ? Whoever responds ? Choose unit of analysis ? Whoever volunteers ? Choose sampling frame ? Whoever is handy ? Identify steps (e.g., college sophomores) ? Zimbardo’s sample? ? Danger: limits generalizability 23 24 4 Design a sampling strategy for ? The effect of premarital counseling on couple communication skills Complexities in the Concept of Cause ? Which is the cause in the causal chain? ? The effect of welfare reform on family ? Causes as inferences self-sufficiency ? The effect of home visiting infants on parenting skills ? The relationship between gender and wages in the human service or hospitality industry ? Validated causal propositions escape disconfirmation ? NEVER confirmed! 25 Causal evidence What’s wrong with only measuring hypothesized cause? Requirements for cause: ? Cause precedes effect ? What happens if alternate cause covaries with hypothesized cause? Implications for a cross -sectional study? ? Presence of effect does not reflect a chance deviation ? Result does not occur without cause ? Congruence: larger change in IV ? larger change in DV ? Control for alternate cause ? Effect follows a detailed prediction ? Rival explanations eliminated Be sure to review these for Exam I! Effect Cause 1 Alternate cause 27 Four cells needed to separate cause from alternate cause Cause Alternate cause High 26 28 Common confounded variables ? Education & income High Low Effect High Effect Low ? Marital status & income ? Ethnicity & income/education ? Interest & experience ? Ability & experience Low Effect High Effect Low 29 30 5 Alternative explanations? Controls? Chapter 9 Other Criteria for Research ? Childhood poverty leads to lower levels of academic achievement ? Active listening leads to increased marital satisfaction ? Enhanced parental involvement leads to enhanced academic performance ? Earlier parent-child separation leads to insecure adolescent attachment 31 32 Criteria to be optimized Audience Credibility ? Audience credibility ? Validity is subjective ? Plausibility of conclusion is critical ? Was hypnotic suggestion a credible strategy to induce deafness in Zimbardo’s study? ? Plausibility lies in eye of beholder ? Avoid having a bias or interest in the ? Balance of internal and external validity outcome ? Resource allocation 33 34 Relative Weighting of Building Audience Credibility ? ? Build on already accepted knowledge Internal Validity (LP) ? ? Avoid weaknesses of previous studies ? Use accepted techniques ? ? Justify any use of the nonstandard External Validity (GP) ? ? E.g., Zimbardo’s use of hypnotically induced deafness ? 35 Does the evidence in the study support the existence of a relationship between or among its variables? Does the relationship between/among the variables in a study generalize beyond the specific characteristics of the study in which it was found? N.B.: Don’t confuse with test validity!!! 36 6 Threats to internal validity Balancing Internal and External Validity ? Control strengthens internal validity; ? Artificiality threatens external validity ? Strengthen internal validity by narrowing scope ? How is internal validity strengthened by conducting research in? ? ? ? Studying infant attachment ? Strange Situation Behavior Instrument science of strange behavior of children in strange situations with strange adults for the briefest possible time.” ? “the ISU CD Laboratory School How is external validity threatened by conducting research in? ? ISU MFT Clinic 37 Resource Allocation 38 Limits and constraints ? Personnel ? Ethical standards ? Travel ? Instruments ? Institutional Constraints ? Resource Limits ? Equipment ? Office ? Telephone ? Copy machine ? Postage 39 Ethical standards ? 40 Institutional Constraints ? First do no harm…. ? ? Access to participants ? Questions permitted ? IRB approval ? Participant informed consent ? Parent informed consent ? Time and location How much time will each person consent to be interviewed? ? How long a survey will each respondent complete without tossing it in the wastebasket? ? Longer survey ? ? Confidentiality Mandated reporter? ? More on ethics on February 25; complete ISU training on the web: ? http://www.research.iastate.edu/training/ Institutional Review Board Human Subjects Committee ? Informed consent ? ? Review by IRB ? Tuskegee study on syphilis ? Makes research on vulnerable subjects difficult because of limited access ? ? ? 41 ? higher refusal (attrition) rate less representative sample! Conclusions permitted???? 42 7 Resource Limits ? Available staff ? Equipment and/or Instruments ? Money ? Travel, postage, phone for administration ? Time for data entry ? Time for data analysis ? Time 43 8