Health Research: Chapter 14
About this set
Created by:
nmr123321 on April 27, 2011
Subjects:
health science, health research, health studies
Description:
Preparation of quantitative data for coding and analysis
Log in to favorite or report as inappropriate.
Order by
26 terms
Terms | Definitions |
|---|---|
 Coding | -method of conceptualizing research data and classifying them into meaningful and relevant categories for the participants in the study -a number is assigned to the category which is called the code -coding of quantitative data should be carried out at the end of the data collection period, once all the codes have been finalized -basic rules for the development of the coding scheme (coding frame) for quantitative data are that the codes must be mutually exclusive and must be applied constantly -sometimes can be handcoded during interview or in teh field, can do it on computers using electronic versions of the questionnaire |
| deductive coding | -if previous knowledge was used to construct response categories before the instrument was administered to respondents then it is called deductive coding |
| Inductive coding | -when there is no previous knowledge it is called inductive coding -advantage is richness of the coes, flexibility and develop new categories that might otherwise have been thought of -disadvantage is time consuming |
| Coding Boxes | -allocated for each question, each box can only contain 1 number -eg. 40 would have 2 boxes |
| Coding transfer sheets | -used if the investigator doesn't want to clutter the questionnaire with numerical codes and coding boxes |
| Code Book | -used for quantitative data -a master copy of the questionnaire with the question number the question, the full range of codes, the number of columns for each response, variable label |
| Numerical values for codes | -qualitative data has to be given numerical value -can be ad hoc or conventional (this one allows for comparison because other studies also use it) -codes have to be sensible and consistent |
| Coding open questions | -have to be grouped by theme and coded -might be given more than a code to ensure that all items mentioned have been accounted for -might simple code for the nature of response (positive, neutral or negative) -application of statistics is limited |
| Coding closed questions | -require that any groupings should be defined before the data are collected -response is then allocated to the pre-defined category with a number assigned -important at the design stage of the questionnaire to ensure that a recorded response is always requested |
| Checking returned questionnaires | next |
| Verification | -2 coders should be used to code the entire set independently and someone should check both results |
| Data entry onto computer | -optical scanners can be used to code directly -verification involves 2 people independently entering data -human coding and entry of direct data entry is preferred. |
| Cleaning the data | -need to eliminate the more obvious errors that will have occurred during the data collection, coding and input stages |
| Range Checks | -for data fields containing information about a continuous variable, observations should within a specified range |
| Consistency Checks | -certain combinations of within-range values of different variables are either logically impossible or very unlikely(children in geriatric ward) |
| Checking for bias in the analyses | next |
| Response bias | -information should be collected about non-responders in order that the difference between the responders and non-responders can be analyzed AGE BIAS? |
| Interviewer Bias? | -occurs when more than one interviewer or observer used |
| Missing values and data checks | -2 types of missing values 1.where a question is deliberately blank because it did not apply to the individual respondent 2.where a reply was expected but not given |
| 9 or 99 or 999 | -used 9 for inadequate responses when coding or 99 or 999 etc |
| 8 or 88 or 888 | -used 8 for "does not apply" responses when coding or 88 or 88 etc |
| Computer packages for the analysis of quantitative data | -for straightforward analysis SPSS used -hierarchical data easier to analyze using other programs |
| univariate statistics | descriptive stats for the analysis or description of one variable |
| bivariate statistics | (descriptive stats for the analysis of the association between 2 variables) |
| -multivariable statistics | (techniques which allow for the measurement of the effects of one variable on an outcome to be measured, while controlling for the effects of other variables, thus removing their effects) |
| Stages in the analysis | -descriptive stats should be first in order to be able to describe the findings and look at any skewness -results of bivariate analyses will inform the investigator about whether multivariable analyses should be performed with any of the variables, depending on the type of data and the aims of the study -trends are then examined |
First Time Here?
Welcome to Quizlet, a fun, free place to study. Try these flashcards, find others to study, or make your own.