Field Research Designs Purpose of field research designs Types of field studies Planning a field study Sampling plan Questionnaire design Online Questionnaires Data analysis Special concerns in field research
Some Definitions Population Element Sample Subject Sampling An entire group of people, events or things of interest Element Single member of population Sample Subgroup of the population Subject Single member of sample Sampling Selecting sufficient number of elements from population so that features of the sample (e.g., mean) can be generalized to the population
Advantages of Sampling Less cost Compared to cost of studying population Less error In collecting & analysing data Less time Because fewer elements considered Less intrusive/destructive E.g., when measurement changes phenomenon
Sampling Plan Random sampling Each member of population has EQUAL chance of being selected into sample Ease of identifying population Company vs. community vs. student populations Determining sample Random number tables, computerized routine, drawing from urn
Sampling Plan Random sampling Randomly select additional participants if initially selected ones refuse/cannot Order population, sample every xth person – ordering is not related to variable of interest E.g., Immigration checks Use of convenience sample Include variables that assess representativeness of obtained sample If response rate is low e.g., ethnic harassment study
Modified random sampling Sampling Plan Modified random sampling Stratified random sampling Divide population into subgroups & randomly select from subgroups Sub-grouping expected to influence results (e.g., motivational levels in R&D vs. secretarial staff) Used when total sample size is small and number of subgroups is large E.g., visible minorities at Scar campus
Modified random sampling (cont’d) Sampling Plan Modified random sampling (cont’d) Cluster sampling Choose participants based on membership of a group Groups are then chosen to participate in study Stats computed can have large sampling errors E.g., examine units in 4 vs. 30 boxes of a shipment Over-sampling from a subgroup E.g., gays in the org’n Need to weight descriptive stats appropriately
Field Research Designs Sampling plan Questionnaire design Online Questionnaires Data analysis Special concerns in field research
Use existing measures of concepts Questionnaire Design Use existing measures of concepts Comparability Reliability (standardization) Validity
Questionnaire Design Writing Items Comprehensiveness Accuracy E.g., commitment scale Accuracy Maintain respondents’ cooperation & dignity
Questionnaire Design Writing Items Structured vs. Open-ended items Respondent involvement in research Purpose of research Exploratory vs. confirmatory Type of question E.g., When all possibilities are not known/too many Resource availability Time & money for coding & analysing Saks 69-73; Sekaran 238-242
Questionnaire Design Writing Items Use simple, direct, familiar language Be clear & specific (avoid ambiguous items) Use positively and negatively worded items Avoid double-barreled items Avoid Leading questions Avoid loaded questions Ensure applicability to all respondents Avoid recall-dependent items Saks 69-73; Sekaran 238-242
Questionnaire Design Writing Items Minimize Response styles Yea/Nay sayers (acquiescence) Positive vs. negatively worded items Social Desirability Forced choice format Content-specific anchors (e.g., BARS) Items scattered across survey Saks 69-73; Sekaran 238-242
Response options in structured scales Questionnaire Design Response options in structured scales Types of Rating Scales Likert, Semantic Differential, Itemized Rating etc. (p. 197-199 Sekaran) Bimodal responding Using only a portion of the response options Ensure anchors have same meaning to all respondents Use numbers w/verbal descriptors
Response options in structured scales Questionnaire Design Response options in structured scales Identify time frame of phenomenon of interest Optimal number of scale points 5 points is best, fewer results in less variability Instructions Provide examples Participants’ education level Previous exposure to method of data collection e.g., web/email surveys
Response options in structured scales Questionnaire Design Response options in structured scales Sequencing General to specific, easy to difficult Avoid placing positively and negatively worded items tapping into the same dimension near each other Beware of ordering effects Issues with dispersal of items Numbering Attend to data analyses issues Sekaran 242
Response options in structured scales Questionnaire Design Response options in structured scales Layout (appearance) Introduction e.g., Study Information Sheet Organization By sections Personal Data Request sensitive personal data at the end Open-ended questions in the end Conclusion Sekaran 245-249
Ways to Optimize Return Rate Questionnaire Design Pre-testing Survey Readability, item content, ambiguities Ways to Optimize Return Rate Upper management or union support Work time allocated for survey completion Coercion, confidentiality concerns Participants’ belief in value of research Previous experience with HR research 30% rate is common
Optimizing Return Rate Questionnaire Design Optimizing Return Rate Professional appearance For mailed survey: use first class mail & include return postage Send reminders Provide Incentives for responding Keep survey at optimal length Identify characteristics of non-responders to establish representativeness of sample Identify mechanism for clarifying questions
Online Questionnaires Advantages Speed Delivery to participants Completed surveys to researcher Cost efficiency Environmental costs (e.g., paper, ink) Personnel costs (e.g., typing, data entry)
Online Questionnaires Concerns Respondents’ access to computers Establish invariance b/w paper-pencil and computer versions (e.g.,achievement, attitude measures) Ballot stuffing Unique access control numbers Start up costs E.g. survey monkey $20/month Technical difficulties during survey administration Researcher’s control over design interface E.g. survey monkey Employee reactions to online surveys
Preliminary Data Cleaning Data Analysis Preliminary Data Cleaning Use descriptive data to catch errors E.g., means, ranges, standard deviations Coding open-ended responses Analysis & Interpretation Descriptive data Frequencies, means Group comparisons T-tests, ANOVAs Establish relations between variables Correlations, regressions
Special Issues in Field Research Scale reduction Alternatives to shortening existing scales Reduce number of variables Use alternative methods of measurement E.g., peer ratings, archival data etc.
Special Issues in Field Research Percept-percept problem Response bias due to cross-sectional, mono-method measurement of all variables Alternatives to self-report questionnaires E.g., archival, objective data Multiple data collection times E.g., Longitudinal study Dispositional influences E.g., neuroticism
Special Issues in Field Research Survey matching Ensure confidentiality & anonymity Controlling extraneous variables Conceptual understanding Sample characteristics Measurement or control of variables
Special Issues in Field Research Response Variability Dichotomous scales (e.g., y/n responses) Ethics Info re: research objectives Precautions re: anonymity Limit demographic info requested Web/email based surveys Mechanisms for research feedback Implications, planned action, follow up