I am conducting research on the effects of network externalities (NE) on technological adoption, where I have to test whether the chosen NEs have an effect on technological adoption (in effect, if there is a correlation between them) and if yes, which of those have a stronger effect. My independent values for NE are two, one of which is a continuous IV and the other - an categorical IV. My dependent value is a dichotomous value (adoption/no adoption). Therefore, I was wondering what type of statistical test should I conduct for both of the NEs (e.g. phi correlation for the categorical IV and point-biserial correlation for the continuous IV)? Also, how do I compare the effects of the different IVs after that, when one is categorical and the other continuous?
Hi, Eleonora,
Here are four free-to-view articles which discuss a modern statistical method that can simultaneously analyze (multi)categorical and ordered attributes ("dependent measures"), and for which the solution explicitly maximizes predictive accuracy for your sample--which no alternative legacy statistical methodology can accomplish:
- https://odajournal.com/2013/09/19/unconstrained-covariate-adjustment-in-cta/
- https://odajournal.com/2013/09/19/aggregated-vs-referenced-categorical-attributes-in-unioda-and-cta/
- https://odajournal.com/2013/09/19/manual-vs-automated-cta-optimal-preadmission-staging-for-inpatient-mortality-from-pneumocystis-cariini-pneumonia/
- https://odajournal.files.wordpress.com/2019/01/v6a8.pdf
Many other examples are available: https://ODAJournal.com
And, citations in these articles identify many other articles on this subject which are found in pay-to-view journals.
Logistic regression analysis (LRA) works POORLY.
(Below, when possible, free-to-view articles are cited--otherwise pay-for-view articles are cited, which require access to an academic library, or payment, to view.)
LRA remains an often-used legacy analysis, however it is increasingly realized that it has many significant shortcomings--NONE of which are true for modern maximum-accuracy methods--commonly referred to as the optimal data analysis (ODA) paradigm.
- https://odajournal.com/2014/04/06/a-statistical-guide-for-the-ethically-perplexed-chapter-4-panter-sterba-handbook-of-ethics-in-quantitative-methodology-routledge-2011-clarifying-disorientation-regarding/
Tremendous shortcomings of LRA include, for example:
LRA identifies effects which in reality DO NOT exist
- https://odajournal.com/2019/05/16/logistic-discriminant-analysis-and-structural-equation-modeling-both-identify-effects-in-random-data/
- https://odajournal.com/2019/04/11/effect-of-sample-size-on-discovery-of-relationships-in-random-data-by-classification-algorithms/
- https://odajournal.com/2019/03/21/some-machine-learning-algorithms-find-relationships-between-variables-when-none-exist-cta-doesnt/
LRA fails to identify effects which in reality DO exist
- Yarnold, P.R., Soltysik, R.C., McCormick, W.C., Burns, R., Lin, E.H.B., Bush, T., & Martin, G.J. (1995). Application of multivariable optimal discriminant analysis in general internal medicine. Journal of General Internal Medicine, 10, 601-606.
LRA does NOT maximize classification accuracy, and LRA models may be misspecified
- https://odajournal.com/2015/12/01/unioda-vs-logistic-regression-and-fishers-linear-discriminant-analysis-modeling-10-year-population-change/
- Yarnold, P.R., Hart, L.A., & Soltysik, R.C. (1994). Optimizing the classification performance of logistic regression and Fisher’s discriminant analyses. Educational and Psychological Measurement, 54, 73-85.
- Yarnold, P.R., & Soltysik, R.C. (1991). Refining two-group multivariable classification models using univariate optimal discriminant analysis. Decision Sciences, 22, 1158-1164.
- Linden, A., & Yarnold, P.R. (2017). Using classification tree analysis to generate propensity score weights. Journal of Evaluation in Clinical Practice, 23, 703-712. DOI: 10.1111/jep.12744
- Linden, A., & Yarnold, P.R. (2016). Combining machine learning and propensity score weighting to estimate causal effects in multivalued treatments. Journal of Evaluation in Clinical Practice, 22, 875-885. DOI: 10.1111/jep.12610
- Linden, A., & Yarnold, P.R. (2016). Using data mining techniques to characterize participation in observational studies. Journal of Evaluation in Clinical Practice, 22, 839-847. DOI: 10.1111/jep.12515
LRA can report invalid p values due to violated underlying assumptions
- https://odajournal.com/2014/03/29/unioda-vs-logistic-regression-serum-cholesterol-and-coronary-heart-disease-and-mortality-among-middle-aged-diabetic-men/
LRA has difficulties with multicategorical variables
- https://odajournal.com/2013/11/11/univariate-and-multivariate-analysis-of-categorical-attributes-with-many-response-categories/
- https://odajournal.com/2013/09/19/aggregated-vs-referenced-categorical-attributes-in-unioda-and-cta/
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis