Chapter 2 Cross-Cultural Research Methods PDF

Preview

Summary

Download Chapter 2 Cross-Cultural Research Methods PDF

Description

Chapter 2: Cross-Cultural Research Methods Types of Cross-Cultural Research - Progressed through dif. Stages; 1st stage had initial tests of cultural difs. & discovery of cultural difs. - ↳ 2nd stage → look for meaningful dimensions of cultural variability that could explain those difs. (ind. Vs. coll. Came from this stage) - ↳ 3rd stage → conceptual application of prev. Dimensions in cross cultural studies - ↳ 4th stage → field is currently in, applying dimensions to document effects Method Validation Studies - Validity: the degree that a finding, measure or statistic is accurate or reps. What it’s supposed to - Reliability: degree that a finding, measure or stat is consistent - Can’t use a measure that was valid in one culture & use it on another → just translating a measure doesn’t mean it’s equal - Cross-Cultural Validation Study: study that examines whether a measure of psych construct that was generated in one culture is applicable/equivalent in another culture - ↳ imp. To conduct b/f cross-cultural comparisons Indigenous Cultural Studies - Indigenous Cultural Studies: studies that use complex descriptions of cultures/cultural difs. To predict test for difs. In psych. Variable - Psych process/beh. Can only be understood in context of culture that it occurs in, so understanding mental processes requires analysis of cultural systems that produce/support those processes - Since it’s indepth analysis of 1 culture, not cross cultural really → insights are compared to similar studies in other cultures so cross cultural after the fact - Type of work roots in anthro, used to explain cultural difs. In psych process like morality, attributional style, nature of unspoken thoughts, etc. Cross-Cultural Comparisons - Cross-Cultural Comparisons: study that compares 2 or more cultures on psychological variable of interest, often w/ hypothesis that one will have significantly higher scores than the other (backbone of cross-cultural research) Types of Cross-Cultural Comparisons Exploratory vs. Hypothesis Testing - Exploratory Studies: designed to examine existence of cross-cultural similarities or difs.; simple, quasi-experimental designs - ↳ main strength is broad scope for identifying cross-cultural similarities/difs. Esp. in under researched areas - ↳ weakness is limited capability to address the cause of difs. (hypothesis testing studies are vice versa) - Hypothesis-Testing Studies: designed to test why cultural difs. Exist; include context variables or use experiments - ↳ leads to more substantial changes in theory development but less likely to discover interesting difs. Outside tested theory/vulnerable to problems related to cultural bias Presence or Absence of Contextual Factors - Contextual Factors: any variable that can explain observed cross-cultural difs.; may involve characteristics of participants (SES, edu., & age) or cultures (eco. developments /religious institutions) - ↳ hypothesis testing needs to include contextual variables

Structure vs. Level Oriented - Structure-Oriented Studies: studies that examine if constructs are conceptualized the same way across cultures (e.g. is depression conceptualized the same way across cultures), their structures (can it be assessed in the same way?), the relationship of the construct to other constructs (is depression & anxiety pos. Related in all cultures?) or the measurement of a construct - Level-Oriented Studies: studies that examine cultural difs. In mean levels of variables (do ppl from dif. Cultures show same level of depression?) - SOS focus on relationships among variables to attempt to identify similarities/difs. - LOS focus on ppl have dif. Amounts of dif. Variables Individual vs. Ecological (Cultural) Level - Ecological (Cultural) Level Studies: cultures/countries are unit of analysis (instead of ind.) - Data from ind. Avg. for each culture - Data fil w/ ind. Level datahave ind. Data rows & variables in columns; eco. Level has culture in rows & variables in columns - ↳ culture data from compiled mean score → imp. b/c they allow examination of psych variables w/ eco. Level variables like climate, pop. density or GNP (link psych variables w/ eco. variables) - Hofstede’s seminal work → 72 countries w/ 117,00 employees of multinational business org. → created the 4 dimensions (ind. Vs. collectivism, power distance, uncertainty avoidance & masc. Vs. fem) - ↳ added a 5th dimension → long vs. short term orientation - Multi-level Studies: involve data collection at multiple levels of analysis, like ind. Level, context, community & national culture - ↳ e.g. how ind. Perform on cog. Task (level 1) may be related to personality traits of those ind. (level 2) & how personality traits relate to cultural values or other variables (level 3) ~~~~~~~Designing Cross-cultural Comparative Research Getting the Right Research Question - Any consideration of research designs start w/ comprehensive/functional knowledge of literature to know gaps in existing knowledge - Field heavy w/ info knowing 2 cultures are dif., struggle now trying to understand why they are dif. - Culture not necessarily produced in top-down fashion; could be ind. Level psych processes that cause beh. Which creates culture in bottom-up way Designs that Establish Linkages between Culture & Psychological Variables - Linkage Studies: studies that attempt to measure aspect of culture theoretically hypothesized to produce cultural difs. & then empirically link that measured aspect of culture w/ dependent variable of interest Unpackaging Studies - Unpackaging Studies: studies that unpackage contents of global, unspecific concept of culture into specific, measurable psychological constructs & examine their contribution to cultural difs. - Context Variables: variables that operationalize aspects of culture that researchers think produce difs. In psych variables (these variables measured in unpackaging studies) Individual-Level Measures of Culture - Individual-Level Measures of Culture: assess variable on ind. Level that is thought to be product of culture - ↳ often used as context variables to make sure that samples in dif. Cultures

actually have the characteristics thought to differentiate them - Most common dimension used w/ this is ind. Vs. collectivism → Triandis said it was cause of many cultural difs. - ↳ individual level IC acts as syndrome that includes values, beliefs, attitudes & beh. - Multi-method approach used ratings of social content of self, perceptions of homogeneity, attitude & value ratings, etc. - Idiocentrism: individualism on ind. Level; how ind. May act in accordance w/ indi. Cultural framework - Allocentrism: collectivism on ind. Level; how ind. Act in accordance w/ collectivistic cultural frameworks - Vertical & horizontal collectivism & individualism - HC = members of ingroup on same level - VC = members of ingroup categorized by hierarchical/status - VI = ind. Are autonomous & unequal - HI = ind. Are autonomous & equal - IC Interpersonal Assessment inventory (assess context specific IC tendencies in interpersonal situations) - ↳ how japanese & american culture interpret facial expressions → differed strongly; linked w/ difs. In ind.-level IC & accounted for cultural difs. In ind.-level IC Self-Construal Scales - Some thought ind. Vs. collectivistic cultures differed in self-concepts they fostered (ind. Supporting independent self-construals & collectivistic supporting interdependent - ↳ development of scales measuring independence & interdependence on ind. Level Personality - Personality used as context variable b/c associated w/ psych processes → cultural difs. Can be explained by dif. Levels of personality traits in each culture Cultural Practices - Things like child-rearing, nature of interpersonal relationships or worldviews Experiments - Experiments: studies where researchers create conditions to establish cause-effect relationships (create conditions/groups but can’t create cultural groups) - ↳ 2 dif. Types of experiments in cross-cultural research (priming/beh.) Priming Studies - Manipulating mindset of participants & measuring the resulting changes in beh. (manipulate mindsets related to culture to see if they beh. Dif. b/c of primed change) - ↳ if there is change, infer that primed cultural mindset caused difs. In beh. (proving link b/w cultural product & psych process) Behavioral Studies - Manipulations of actual environments & changes in beh. As function of environment → collectivistic cultures cooperate only b/c cooperation is necessary for groups to function effectively Bias & Equivalence - Bias: difs. That don’t have exactly same meaning w/i/across cultures; lack of equivalence - Equivalence: state or condition of similarity in conceptual meaning & empirical method b/w cultures → allows meaningful comparisons - ↳ greater bias is in study, less comparison it has → only if framework & hypotheses have equal meaning in both cultures (& methods of data collection,

management & analysis) Conceptual Bias - Conceptual Bias: degree that theory or set of hypotheses being compared across cultures are equivalent - ↳ testing something a certain way that isn’t applicable across cultures → is theory created w/i western european cultural framework meaningful to another culture? - Intelligence thought to be verbal/analytical critical-thinking skills (WAIS) for IQ → other cultures consider nobility of character as markers of intelligence Method Bias Sampling Bias - 2 issues if CC samples can be compared: - 1. Is sample is rep. Of culture → mostly use uni students so mexican & americans use 2 dif. Universities from both, so is that one uni rep. For all american culture - 2. Are samples equivalent on noncultural demographic variables (age, sex, religion, etc.) - 2 groups from america vs. india; dif. SES, ages, edu.,etc → are these groups really comparable/are difs. Caused only by culture → need to control demographic variables in 2 possible ways: - ↳ 1. Experimentally controlling them by holding them constant in group - ↳ 2. Statistically controlling them when analyzing data - Some factors can’t be held constant → religion is dif. In meaning & practice across culture (i.e. being catholic in US is dif. Than in Japan) Linguistic Bias - Linguistic Bias: semantic equivalence b/w protocols (instruments, instructions, questionnaires) used in CC comparison study - 2 methods used to get linguistic equivalence: - 1. Back Translation: technique of translating research protocols, taking it from one lang., translating it to target lang. & having someone else translate it back to the original → if the same then are equivalent; procedure is repeated if not - Decenter: concept underlying back translation that involves getting rid of culture specific concepts of original lang. Or translating them equivalently into target lang. - 2. Committee approach → several ppl translate protocol into target lang. & debate dif. Forms, words & phrases that could be used (shared consensus of linguistically equivalent) - Can use both protocols Procedural Bias - Some uni students are encouraged to participate, some are forced for grades → expectations about experience w/ research participation differs - ↳ laboratory vs. field, night vs. day, questionnaire vs. observation all have dif. Implications depending on the culture Measurement Bias - Measurement Bias: degree that measures used to collect data in dif. Cultures are equally valid/reliable - Linguistic equivalence doesn’t = measurement equivalence → just because 2 words being used are the same doesn’t mean that they have same nuances - Operationalization: ways researchers conceptually define a variable & measure it - Psychometric Equivalence: degree that dif. Measures used in CC comparison study are

statistically equivalent in cultures being compared - ↳ if questionnaires have the same structures → Factor Analysis: statistical technique that allows researchers to group items on questionnaire; theoretical model underlying factor analysis is that groups of items on questionnaire answered in similar ways b/c they are assessing same underlying psych construct - Questionnaires across cultures → if same group of items would emerge in dif. Cultures → Structural Equivalence: degree that measure used in CCS produces same factor analysis results in dif. Countries being compared (if not, measure is biased) - Internal Reliability: degree that different items in questionnaire are related to each other & give consistent responses - If measuring same mental constructs, items should be related to each other Response Bias - Response Bias: systematic tendency to respond in certain ways to items or scales → dif. To compare if “true” difs. Are being measured - Socially Desirable Responding: tendencies to give answers on questionnaires that make oneself look good - ↳ Self-deceptive enhancement → seeing oneself pos. - ↳ impression management → american vs. korean → american score higher on self deceptive(individualistic) but the latter score high on impression management - Acquiescence Bias: tendency to agree rather than disagree w/ items on questionnaires → countries near mediteranian showed more of both - Individualistic cultures less likely b/c maintaining harmony less stressed than in collectivistic - Extreme Response Bias: tendency to use ends of scale regardless of item content - More masculine culture (shows decisiveness; valued in those cultures) - Reference Group Effect: idea that ppl make implicit social comparisons w/ others when making ratings on scales; ppl’s ratings will be influ. By implicit comparisons they make b/w themselves & others & these influ. make comparing responses across cultures - Japanese responses are more individualistic b/c the internally compare themselves to other collectivistic ppl so inflate the answers Interpretational Bias Analyzing Data - Use chi-square or ANOVA → null hypothesis testing → compare difs. One would expect based of chance alone & probability that it is obtained solely by chance (less than 5%, than findings probably didn’t occur based on pure chance) - Used to see “significant findings” and assume it’s meaningful to most in that culture & is applicable - ↳ used to make broad, sweeping statements that’s not necessarily true → cultural effect size stats Dealing With Non-Equivalent Data - Almost impossible to create entirely equivalent methods → use best approximations of closes equivalents in theory & methods - 4 ways to deal w/ it; - 1. Preclude Comparison - not make comparison in first place ~~~~~~~~~ - 2. Reduce the nonequivalence in the data - take steps to identify equivalent & nonequivalent parts of methods & refous comparisons on only equivalent parts - 3. Interpret the nonequivalence - interpret non-equivalence as imp. Piece of info concerning cultural difs. - 4. Ignore it - don’t do anything - Lack of equivalence means more interpretation regarding their findings

Interpreting Findings - Data they find put through own cultural filters & biases can affect interpretation → jap. Vs. american → jap less expressive but actually americans exaggerate emo. Responses - quasi -experimental, correlational studies → only correlational inferences can be found - Cultural Attribution Fallacies: mistaken interpretation in CC comparison studies; when researchers infer that something cultural produced difs. They observed in their study, despite the fact that it’s not empirically justified in doing so b/c they didn’t measure those cultural factors ~~~~~~~...