Report

Nursing Research 63-377 Dr. Wally J. Bartfay “The most important lesson that can be given to nurses is to teach them what to observe, how to observe, what symptoms indicate improvement, what the reserve…” Florence Nightingale (1860) Methods of Statistical Data Analysis for Quantitative Nursing Research (1) PARAMETRIC TESTS: must have following 3 assumptions present (a) sample has been drawn from a population for which the variance can be calculated, & distribution is expected to be normal (or near normal) (b) level of measurement should be interval or ordinal with an approximately normal distribution (c) data can be treated as random samples Methods of Statistical Data Analysis for Quantitative Nursing Research (2) NONPARAMETRIC TESTS: Statistical techniques used when the assumptions of parametric statistics are NOT met Most commonly used for nominal & ordinal level data Methods of Statistical Data Analysis for Quantitative Nursing Research: (1) NOMINAL DATA: (Independent Groups, 2 variables) (a) Chi-Square: used to determine significant differences between observed frequencies within the data & frequencies that were expected (b) Phi coefficient: used to determine relationships in dichotomous, nominal data (c) Cramer’s V: Analysis technique that is a modification of Phi for continency tables larger than 2 X 2 (d) Contingency Coefficient: a stat. test used with 2 nominal variables & is the most commonly used of the chi-squaredbased measures of association Methods of Statistical Data Analysis for Quantitative Nursing Research (2) ORDINAL DATA (Independent Groups, 2 variables) : (a) Mann-Whitney U: has 95% of the power of t-test to detect differences between groups of normally distributed populations (b) Spearman’s Rank-Order Correlation: A nonparametric analysis technique that is an adaption of Pearson product moment correlation used to examine relationships among variables in a study (c) Kendall’s Tau: Nonparametric test to determine correlation among variables used when both variables are at the ordinal level Methods of Statistical Data Analysis for Quantitative Nursing Research (3) RATIO-LEVEL and/ or INTERVAL DATA: (Independent Groups, 2 variables) : (a) T-test for independent samples: Parametric analysis tech. used to determine significant differences between measures of 2 samples (b) Pearson’s Correlation: Parametric tests used to determine the relationship between variables (-1 to +1 = perfect correlation, whereas 0 = no correlation) (c) ANOVA: Used to examine differences among 2 or more groups by comparing the variability between the groups with the variability within the groups (d) Simple linear regression: parametric analysis tech. that provides a means to estimate the value of a dependent variablebased on the value of an independent variable Bivariate Statistical Data Analysis (1) Nominal data: (a) McNemar Test (2 samples): nonparametric tests in which a 2 X 2 table is used to analyze changes that occur in dichotomous variables (b) Cochran Q test (3 or more samples): nonparametric test that is an extension of above for two related samples Bivariate Statistical Data Analysis (2) ORDINAL DATA: (a) Sign test (2 samples): a nonparametric tech. developed for data that is difficult to assign numberical values to but the data can be ranked on some dimension (b) Wilcoxon matched paired signed ranks tests (2 samples): Nonparametric tech. used to examine changes that occur in pretest/posttest measures or matched-pairs measures (c)Friedman Two-way analysis of variance by ranks (3 or more samples): Nonparametric test used with matched samples or in repeated measures Bivariate Statistical Data Analysis (3) Interval or Ratio Data: (a) T-test for related samples (2 variables): a parametric test used for analyzing the difference between two means (b) Analysis of covariance (ANCOVA)(3 or more samples): Used to test the effect of one or more treatments on different groups while controlling for one or more extraneous variables (covariates) Multivariate data analysis Interval or Ratio Data (3 or more variables): (a) Multiple regression analysis: extension of simple linear regression with more than one independent variable entered (b) Factorial analysis of variance: mathematically it is a specialized form of multiple regression- a No. of types of factorial ANOVA’s have been developed to analyze data from experimental designs (c)Structured equation modeling: designed to test theories (d) Time-series analysis: tech. designed to analyze changes in a variable across time and thus to uncover a pattern in the data Type I Errors occurs when researcher concludes that the samples tested are from different populations (difference between groups is significance), when in fact the samples are from the same population (the difference between groups is NOT significant) Null hypothesis is rejected when it is, in fact, true Can help to decrease chance of this, in part, by increased significance level (e.g., from p < 0.05 to p < 0.001) Type II errors Occurs when the researcher concludes that there are no significant difference between the samples examined when, in fact, a difference exists Null hypothesis is regarded as true, when it is false Can help to control for this, in part, by increasing sample size, which increases power of study (do power statistical power analysis prior to determine the risk of type II and needed sample size) Errors in Data Collection in Qualitative Research Can be introduced in 3 major ways: (1) In the process of data collection; (2) With the process of data analysis, and (3) When both of the above occurs Errors in Data Collection in Qualitative Research Qualitative researchers employ several techniques and processes to ensure the rigor of the data collection and analysis processes (1) RIGOR: term reflects overall quality of study It is reflected in the consistency of data analysis and interpretation, trustworthiness of the data collected, transferability of the themes & credibility of data is ensured Errors in Data Collection in Qualitative Research (2) Trustworthiness: refers to the “honesty” of the data collected Hence, researcher must have meaningful relationship with subjects & have willingness to share their feelings, insights, experiences Consistency in the data collection process helps to establish this Errors in Data Collection in Qualitative Research (3) Confirmability: refers to the consistency and repeatability of the decision making about the process of data collection & analysis Can do “audit trail” which is an ongoing documentation regarding the researcher’s decisions made by researcher (e.g., field notes about processes undertaken) Errors in Data Collection in Qualitative Research (4) Transferability: refers to the extent to which the findings of a study are confirmed by or are applicable to a different group or in a different setting from where the data was collected Different from “generalizability” B/C focus is NOT on predicting specific outcomes in a general population Rather, focus is on confirming that what was meaningful in one specific setting or group, is also meaningful & accurate in a different setting or group (e.g., caregiver burden experienced by stroke survivors with aphasia versus those without aphasia) Errors in Data Collection in Qualitative Research (5) Credibility: overlaps with transferability & trustworthiness- refers to the confidence that the researcher & user of research can have in the truth of the findings of the study Can do “member checks” where data & findings are brought back to subjects to seek their input, check for accuracy, completeness & interpretation of data Can also do “Triangulation”: a process of using more than one approach or source to include different views or to look at the phenomenon from different angles (e.g., Q-sort methods & structured interviews) Analysis of Qualitative Data: General overview Various methods have been developed (e.g., Glaser and Strauss/ Corbin’s methods; phenomenologic methods etc) All involve coding of data schemes and “clustering” of emerging themes from the data (thematic analysis of data) Qualitative Data Analysis Computer programs available now to help researchers organize & analyze their data (e.g., NU.DIST; N. VIVO;ATLAS/TI; MARTIN; QUALPRO and HyperQual2) Importantly, these programs DO NOT perform the thinking & conceptualizing that is at the heart of qualitative data analysis Can help to organize data around “themes” as they are identified in the data Have a fantastic week. Come and see me regarding your research proposals…