Explore chapters and articles related to this topic
Methodological Concerns in Traumatic Brain Injury
Published in Mark R. Lovell, Ruben J. Echemendia, Jeffrey T. Barth, Michael W. Collins, Traumatic Brain Injury in Sports, 2020
Stephen N. Macciocchi, Jeffrey T. Barth
In addition to placing an emphasis on effect size as well as statistical significance, researchers should consider other issues. For example, most neuropsychological studies employ multiple dependent measures. In such cases, corrections for the number of statistical tests undertaken must be employed. Some authors have used Bonferroni or other corrections which statistically holds alpha at the traditional .05. Statistical procedures such as Multivariate Analysis of Variance (MANOVA) are used to keep alpha levels at a nominal level, but when using a MANOVA, researchers should only interpret individual ANOVA’s when the multivariate test is significant. Even when the overall MANOVA is significant, separate ANOVA’s do not take into account the correlation between dependent measures (Weinfort, 1998). In any case, as the number of statistical tests increases, the probability of finding statistically significant differences between experimental and control groups by chance increases substantially. Actually, if six dependent measures are used without a correction procedure the overall alpha rises to .265 which is quite higher than the traditional p<.05 (Weinfort, 1998). Consequently, statistical conservatism should be considered in concussive research, particularly when numerous dependent measures are employed.
An MMPI Study of Females With Allergic Rhinitis: Prediction of Type I Allergic Disorders
Published in Alan J. Husband, Psychoimmunology CNS-Immune Interactions, 2019
M. Gauci, A.J. Husband, H. Saxarra, M.G. King
A between-subjects multivariate analysis of variance (MANOVA) performed on AR scores (long and short forms) of allergic and nonallergic subjects revealed significant differences in mean scores between groups (F[2,51]=11.41, p <0.001). Univariate F tests revealed that significant differences occurred between allergic and nonallergic females for both the long (mean±SE=8.09+0.67 and 4.41±0.51, F[1,52]=17.83, p<0.001) and the short form of the AR Scale (mean±SE=4.50±0.38 and 2.05±0.34, F[1,52]=23.03, P <0.001).
Treatment Comparisons in Clinical Trials
Published in Ding-Geng (Din) Chen, Karl E. Peace, Pinggao Zhang, Clinical Trial Data Analysis Using R and SAS, 2017
Ding-Geng (Din) Chen, Karl E. Peace, Pinggao Zhang
Multivariate analysis of variance (MANOVA) is simply an ANOVA described in Section 3.2.1.2 with multiple dependent variables such as in the Table 3.1 for diastolic blood pressure (DBP) measures of “DBP1”, “DBP2”, “DBP3”, “DBP4” and “DBP5”. As seen in Section 3.2.1.2, ANOVA tests for the difference in means between two or more groups. Therefore, MANOVA extends the ANOVA for univariate to test the difference in two or more vectors of means from these multivariate dependent variables with special consideration of the dependence where a “covariance” structure is included in the MANOVA. MANOVA can then take these correlations into account when performing the significance test. Testing the multiple dependent variables is accomplished by creating new dependent variables that maximize group differences. These artificial dependent variables are linear combinations of the measured dependent variables. For more complete description of MANOVA, readers can refer to Hand and Taylor (1987), Krzanowski (1998) and Anderson (1994).
The Role of Self-Determined Motivation and the Potential for Pre-registration Student Learning: A Comparative Study within a Mental Health Clinical Placement
Published in Issues in Mental Health Nursing, 2023
Christopher Patterson, Michelle Roberts, Dana Perlman, Lorna Moxham
All data were exported from Qualtrics into SPSS and scanned to identify any missing participant entries. This screening identified 23 incomplete surveys which were omitted from the study. Participant data (N = 256) was then condensed (using the procedures identified in the Data Collection Measures section above) into the central research variables of Self-Determined Motivation, Therapeutic Relationship (PC, ED and PCI domains) and MHCC. There were then two main stages to categorise the data for subsequent between-subject comparisons. Firstly, each participant’s self-determined motivation score was ordered from highest to lowest. Next, the participants were classified into one of three “motivation level” groups based on their score in relation to the wider dataset: high (top third of scores), middle (middle third) or low (bottom third). Descriptive and Reliability analyses for the main study variables of TR and MHCC were calculated for each motivation group. To examine the first research question, a multivariate analysis of variance (MANOVA) with Tukey’s Post-Hoc test for PC, ED and PCI was performed. To examine research question two, a univariate ANOVA with a Tukey’s Post-Hoc test for MHCC was used.
A coefficient of discrimination for use with nominal and ordinal regression models
Published in Journal of Applied Statistics, 2021
Thomas J. Smith, David A. Walker, Cornelius M. McKenna
Finally, a third indicator of the variability in predicted probabilities can be computed as η2 coefficient—familiar, of course, as the multivariate η2 measure of effect size in multivariate analysis of variance (MANOVA)—provides an indication of the extent to which the group centroid predicted probabilities vary around the grand centroid. It is defined as the proportion of the generalized variance in the predicted probabilities that is explained by group membership. A large value of η2 indicates that a large proportion of the generalized variation in predicted probabilities is explained by individuals’ actual outcome category membership, and that the predicted probabilities of group membership vary greatly between the groups, reflective of a large degree of discrimination provided by the nominal/ordinal regression functions. For the nominal regression model predicting cell/smart phone ownership, the observed value for this statistic was η2 = 0.220, while for the ordinal regression function predicting how difficult it was for an individual to give up her/his landline telephone, η2 = 0.168. Approximately 22% and 17% of the multivariate generalized variation in predicted probabilities of, respectively, cell/smart phone ownership and difficulty in giving up a landline telephone was explained by an individual’s actual response to the prompts.
Physical activity, nutrition, and psychological well-being among youth with visual impairments and their siblings
Published in Disability and Rehabilitation, 2021
Justin A. Haegele, Xihe Zhu, Patrick B. Wilson, T. N. Kirk, Summer Davis
Data for this study were analyzed descriptively and inferentially. First, means and standard deviations of participants’ age, MVPA, and total difficulty and prosocial scores from Strengths and Difficulties Questionnaire were computed. Following, frequency distribution of the participants’ body weight status, visual impairment classification, and Beverage and Snack Questionnaire categories were reported. The proportion difference in weight status, difficulty and prosocial categories were analyzed using McNemar’s z test. To examine the relationship between participants’ MVPA, BMI, difficulty, and prosocial scores, Pearson correlation analyses were conducted separately for the youth with visual impairment and their siblings. Data normality was examined using P-P plot, equality of covariance using Box’s test (Box’s M = 11.12, F10,7649 = 0.99, p = 0.45) and equality of error variances using Levene’s test (ps ≥ 0.15) for multivariate analysis of variance (MANOVA). Based on the correlation analysis results, a MANOVA was conducted to examine the differences in total difficulty and prosocial scores. Due to the low correlation between MVPA and Strengths and Difficulties Questionnaire scores, a dependent sample t-test (one-tailed) was conducted to examine whether the sibling without visual impairment had significantly higher MVPA than youth with visual impairment. A dependent t-test was used because paired participants (i.e., sibling) from the same families were included in the analysis and MVPA was likely to correlate [44,45]. An α = 0.05 was used throughout the data analyses.